Search interesting materials

Friday, August 12, 2022

Timeliness in government contracting: Evidence from the country's largest metro-rail network

by Anirudh Burman and Pavithra Manivannan.

Introduction

Infrastructure projects in India are plagued by delays (MOSPI, 2022). Proposed explanations include failures of government contracting for public procurement (Singh, 2010, Sinha and Vatsa, 2021, etc.). In this article, we measure delays in the procurement process of the largest metro-rail network in the country -- the Delhi Metro Rail Corporation (DMRC) -- which is considered a successful project. It fares well on global ranks on some parameters such as network length and ridership. The early phases of this metro rail project have been lauded for timeliness in execution and contract payments (Expenditure Management Commission, 2016).

We look at two distinct datasets to obtain a birds eye view and a procurement-oriented view of the delays in DMRC. We find that DMRC is prompt in stages of the contracting processes for which we are able to find evidence, but that the overall project implementation suffers from time overruns. We put this knowledge together to obtain insights into government contracting.

Our approach

As with contracts drawn between any two counterparties, government contracting is a pipeline that runs through four phases (Mehta and Thomas, 2022): (I) Contract specification and design, (II) Contract tendering and award, (III) Contract management and (IV) Contract closure. Flaws in government contracting shows up as inefficiencies in public procurement such as delays in infrastructure projects, which in turn, results in cost overruns, loss in revenues, vendor dissatisfaction and lack of competition when government wants to procure, and ultimately, deprives the public of the intended benefits.

We use three data-sets to understand the timeliness of government contracting in DMRC projects.

  1. A data-set of the time taken by the various DMRC projects, sourced from the CapEx database published by the Centre for Monitoring Indian Economy (CMIE).
  2. A data-set of tenders awarded by DMRC, hand-constructed from the 'Contracts Awarded' section of the DMRC website;
  3. A data-set of payments made by DMRC, hand-constructed from the 'Vendor Payment Details' section of the DMRC website and from an RTI application.

The first data is sourced from the CMIE CapEx database. The CapEx database records the date of significant events for each project. We collect data for the three operational metro networks constructed by DMRC, that is, Phase 1, 2 and 3. This data-set consists of project level information such as, the date of announcement of the project, initial completion date, actual completion date and time overruns.

The second dataset is a hand-constructed data-set consists of tender level information for awarded contracts of DMRC, such as the department calling for tenders, nature of work, date of publication of Notice Inviting Tender (NIT), date of issue of letter of acceptance and value of the contract. This data-set covers this information for 892 tenders for a period of 5 years (2016-2020). DMRC categorises these tenders into 7 heads: Civil and Architecture Works, Electrical Works, Operations and Maintenance, Rolling Stock, Track Works, Signalling and Telecom and Property Development.

In addition, we categorise the contracts for IT services and housekeeping works as 'Miscellaneous' and the procurement done by DMRC for other metros in the country as 'For Other Metros'. The highest number of contracts were awarded for Operation and Maintenance works (623) and the least for Rolling Stock (2). Table 1 shows the typology of procurement undertaken by DMRC during our study period.

Table 1: Typology of DMRC Procurement
Category 2016 2017 2018 2019 2020 Total
Civil and Architecture Works 17 28 9 15 9 78
Electrical Works 5 1 3 5 13 27
Operations and Maintenance 0 65 167 200 191 623
Rolling Stock 0 0 0 2 0 2
Track Works 12 2 0 1 3 18
Signalling and Telecom 8 1 0 0 3 12
Property Development 0 5 3 10 5 23
Miscellaneous 4 14 23 8 13 62
For other metros 5 7 17 8 10 47

Our second hand-constructed data-set consists of monthly bill payment status of DMRC. Pursuant to government communication (vide D.O.18(18)/IFD/2019 dated 05.11.2019), DMRC uploads its monthly vendor payment details on its website since December 2019. This data gives us periodic information about the bill submission date and the bill payment date of DMRC vendors. Since the website does not archive its data we obtained our data partly from the DMRC website and partly vide an RTI application made for this purpose. Our data-set consists of 20,654 bills for the period between November 2019 to August 2021. The payment period is unknown for about 1,550 bills in our data-set, which we discard in our analysis.

We restrict our study to benchmark DMRC's performance against the timelines prescribed by its internal guidelines (DMRC Procurement Manual, 2016 and General Conditions of Contract, 2019) and the Central Government procurement guidelines (General Financial Rules, 2017 and the Manual for Procurement of Works, 2019). We do not employ a comparative analysis with other procuring entities for two reasons: One, availability of data in government portals such as the CPPP (Central Public Procurement Portal) and websites of the procuring entities are often sparse and sporadic. Second, a deeper understanding of the fundamental functioning, internal rules, processes and organisational structure of each entity is required for a meaningful comparison and it warrants a separate study.

Findings: Time overrun in DMRC project implementation

In the CapEx data, we are able to see that, between 1995 to 2021, there were three projects announced, implemented and completed by DMRC. These are the Phase 1, Phase 2, and Phase 3 lines. These have been operational from 2006, 2011 and 2021 respectively. From this data, we are able to locate various timelines for all three phases, including the date on which the Phase was announced, to the date on which they were completed and commissioned for public use for fully operational metro lines. We calculate the time overruns as the difference between the date projected initially as the completion date for a Phase and the date on which it was actually completed and operationalised. These are presented as project delays in Table 2.

Table 2: Project delays (in months)
Phase 1 15
Phase 2 32
Phase 3 102
Source: CMIE Capex Database

Findings: Timeliness in contract award by DMRC

High-income countries, countries with greater political accountability, and countries with greater economic freedom process public works procurement in a more timely manner (Djankov and Bosio, 2020). Each of these countries does infrastructure procurement following its own regulations to award contracts. In India, works procurement is guided by the Manual for Procurement of Works, 2019. According to Clause 5.6 in this manual, the time taken by Ministries and Departments from the date of opening the tender to the date of awarding of contract is 90 days.

We estimate the actual time taken by DMRC to award tenders (Table 3). This is calculated as the time taken from the date of opening of the tender to the date of issuing of the acceptance letter. We find that, on an average, DMRC takes 91-92 days to complete the tendering process.

Table 3: Time taken to award tenders (in days)
Year No. of tenders Average time taken
2016 51 101
2017 123 98
2018 222 103
2019 249 81
2020 247 89
Average 178 92

Findings: Timeliness in making vendor payments by DMRC

Payment delays are endemic in public contracts in India. DMRC has sought to avoid payment delays by including provisions for both interim and final payments within its Procurement Manual and General Conditions of Contract, 2019 (GCC). Depending on the type of contract, payments may be made at different stages of the procurement cycle. At Clause 11, the GCC provides for set timelines for the scrutiny of invoices and payments to be made by the procuring entity:

  • Interim payments: A contracting firm may apply to the respective project engineer of DMRC requesting for an 'interim payment certificate'. This certificate will be issued based on achieved milestones or prescribed payment schedule in the contract, if any.
    1. Within 21 days of the request, the project engineer must issue the interim payment certificate specifying the amount due to the contractor.
    2. DMRC is mandated to make 80% of the certified payment amount within 7 days of issue of the certificate.
    3. The balance 20% is to be made within 28 days of issue of the certificate.
  • Final payments: Once the project engineer certifies that the contractor has completed all his obligations related to a particular work, the contractor is entitled to apply for a 'final payment certificate' with the required supporting documents.
    1. Within 28 days of receiving this request, the project engineer must issue the final payment certificate stating the final amount due.
    2. DMRC is mandated to pay the amount certified in the final payment certificate within 56 days of issue of this certificate.

We look at the vendor payments data-set of DMRC to analyse the adherence to these timelines. We find that, on an average, DMRC takes about 4 days to clear its dues from the date of submission of the bill by the vendors (Table 4). For the data-set in our study, the maximum days taken by DMRC to make its payment is about a year.

Table 4: Time period for clearance of dues (in days)
Year No. of bills Average time Median time Minimum time
2019 1326 5 2 0
2020 9887 4 3 0
2021 7891 4 4 0
Total 19104 4 3 0

Payment delays by public sector enterprises in India to their vendors far exceeds their procurement values (Manivannan and Zaveri, 2021). We find that DMRC is an outlier in terms of maintaining payment discipline to its vendors, and in adhering to the timelines provided in its GCC.

Discussion

The public discourse on government infrastructure procurement focuses on delays and time overruns being an indicator of poor government contracting. In this article, we have analysed the capability of a procurement-intensive public sector enterprise to keep up with its timelines in two stages, that is, in contract award and payments. We find that DMRC takes about 3 months to award a contract, and about 4 days to clear its payment dues to vendors. Regardless of this exemplary performance on awarding contracts and paying vendor dues, we also find that the overall project implementation by DMRC failed to meet scheduled timelines to complete. All three phases took a longer time than originally expected. In fact, we observe that the overall project time delays increased from the Phase 1 project to the Phase 3 project.

Executing infrastructure projects on time has been a continuous concern and challenge in India. Our analysis about DMRC timeliness in awarding contracts and in making payments provides evidence against the popular perception that public projects are delayed due to delays in decision-making by the public authorities and their inability to make timely payments. Instead, we speculate that these are because of other factors for overall project delays, some of which could be misaligned allocation of scope and risk in procurement contracts and poor contract management. A deeper analysis into each project, procurement practises, financial and institutional structure of DMRC may help in understanding the reasons for its timely performance in certain procurement processes and the potential causes of time overruns. These learnings can then be adopted by other procuring entities to achieve better procurement and project outcomes.

References

Ministry of Statistics and Programme Implementation Infrastructure and Project Monitoring Division, 434th Flash Report on Central Sector Projects, January 2022.

Ram Singh, Delays and Cost Overruns in Infrastructure Projects: Extent, Causes and Remedies, Economic and Political Weekly, Vol 45, No. 21, May 2010.

PC Sinha and Ananys Vatsa, Delays in Project Completion in India, Indian Journal of Projects, Infrastructure and Energy Law, January 2021.

Erica Bosio and Simeon Djankov, Timely procurement of public works, World Bank Blogs, February 2020.

Department of Expenditure, Ministry of Finance, General Instructions on Procurement and Project Management, October 2021.

Expenditure Management Commission, Recommendations of the Expenditure Management Commission, December 2015.

Department of Expenditure, Ministry of Finance, Manual for Procurement of Works, 2019.

Delhi Metro Rail Corporation Ltd., General Conditions of Contract, November 2019.

Pavithra Manivannan and Bhargavi Zaveri, How large is the payment delays problem in Indian public procurement?, The Leap Blog, March 2021.

Charmi Mehta and Susan Thomas, Identifying roadblocks in highway contracting: lessons from NHAI litigation , The Leap Blog, July 2022.

Charmi Mehta and Diya Uday, How competitive is bidding in infrastructure public procurement? A study of road and water projects in five Indian states , The Leap Blog, March 2022.


Anirudh Burman is an Associate Research Director and Fellow at Carnegie India. Pavithra Manivannan is a Senior Research Associate at XKDR Forum and Chennai Mathematical Institute. We thank Susan Thomas for valuable comments and discussions.

Wednesday, July 13, 2022

More ammo: Improving resilience against extreme surges in demand

by Ajay Shah.

The Javelin anti-tank guided missile is important for the defence of Ukraine. Under normal times, the production capacity seems to be about 3600 a year. The Ukrainians seem to be using 500 per day, or roughly one missile per kilometre of battlefront per day. The peak load is about 50 times bigger than normal times.

Vershinin, 2022 estimates the Russian army is using 7,176 artillery rounds a day, and argues that these numbers are challenging for the modern Western military manufacturing capacity. He estimates that present US annual artillery production would last for about two weeks of combat in Ukraine. In more recent times there are estimates about Russian use of as much as 60,000 rounds/day.

It may appear that with precision guided weapons, a smaller number of weapons will be required to get the job done. However, precise information about targets is lacking, and the military is reduced to shooting at numerous low probability targets. There are more pathways to target acquisition owing to drones, low earth satellites, night vision, etc., and therefore there are more opportunities to use ammo per unit time. The Ukrainian Armed Forces (UAF) innovated with their new `GIS Art for Artillery' system, where rumoured gains on the delays in the kill loop run from 20 minutes to 30 seconds. As a consequence, modern wars are facing production constraints. As an example, in the small air war in Libya in 2011, the UK and France quickly ran out of precision guided munitions (PGMs).

Such problems with the peak-to-base ratio are not unique to ammo. Consider medical oxygen. The peak load in the delta wave was much bigger than normal times. Alongside this, the bulk of the oxygen production is in the economically advanced peninsula, the biggest demand was in the Hindi heartland, and transporting oxygen is difficult as the refrigerated trucks cannot go at over 25 kph.

Or consider surgical masks and personal protective equipment (PPE). The peak demand during the pandemic perhaps went up by 50 to 100 times when compared with normal times.

Or consider medical education. With students returning from Ukraine, there was a small surge in demand for medical education in India. In a healthy economy, there should be a supply response. In a well functioning society, the resource allocation is not fixed.

Or consider electricity. Electricity demand peaks in the evening, windmills are unreliable, the sun shines in the day and can be obscured by clouds. There is substantial intra-day variation of demand (that is quite predictable), but supply is unpredictable and has a different natural intra-day variation. The puzzle of the energy system lies in dealing with the peak-to-trough ratio.

How should we think about such problems? How does the price system respond to these challenges? Is there market failure? What, if anything, is the role for the state in improving things?

Surge capacity as an option

The right but not the obligation to buy is an option. When the buyer has the right, but not the obligation, to buy 2$\times$ more or 20$\times$ more from the seller, at a preset price, this is an option.

As we know from the field of options, options are always valuable (i.e. they come with a non-zero cost). Being there with excess capacity is not free, for the seller. And, the value of the option goes up when there is more volatility. While financial options loom large in the imagination, the world is full of real options.

Surge capacity in the price system

Prices move, from moment to moment, till supply equals demand. When faced with a shortage, prices go up so as to ration out many prospective buyers. And, equally, those high prices tickle firms into producing more. Vast amounts of patience and intelligence are put in, by buyers and sellers, in order to reduce demand (e.g. by finding substitutes) and increase supply (e.g. by producing in innovative ways). Every surge in prices contains the seeds of its own demise, as buyers establish alternatives, and through the supply surge that follows. Covid vaccines were always going to be a short hot market, and production is now being shut down.

When demand surges or supply drops, the price system sends out signals for firms to produce more through high prices. This tends to be accompanied by a lot of hand-wringing about shortages and high prices. If you think "something should be done", to increase output, you should be happy at what is unleashed by the price system, as there is no force more powerful than high prices, in encouraging buyers to buy less and sellers to produce more.

The market economy is not a bureaucracy; it thinks in all sorts of creative ways. If the price of oxygen is high enough, steel factories will stop making steel and sell oxygen into the public market. If prices go sufficiently high, oxygen cylinders from the Indian peninsula, and from abroad, would be airlifted into the Hindi heartland. The sources of increased supply will always surprise us.

But with the best of effort, mobilising enhanced output is hard and takes time. There is a cost to reallocating the resources of the economy, in order to shift from making widget $x$ to widget $y$. The price system finds this reallocation at the lowest cost to society, at the lowest disruption to society, without harming the incentives for sound behaviour and long-term growth.

Many times, a disruption on the output side is also a disruption of the inputs of the firm. When vast increases in output are required, the inputs (whether physical raw materials or the precise human capital) also become costlier. Both supply and demand curves change in many a surge. Such a combination of factors exacerbates the price rise.

Resilience

The price system will sort things out, in the sense of finding the price at which demand equals supply. It is interesting to go one step further and ask: How big was the supply response, of masks produced per day at its peak divided by masks produced per day before the pandemic? A more resilient economy is one where the price system induces a bigger output surge in a shorter time while requiring a smaller rise in the price.

Alternatively, we can focus on quantities and wonder, under what conditions can very large surges be achieved? We can identify a few sources of resilience.

Complexity in the production process

In a country where many things are produced, and in a country with deep pools of skilled people, there will be more headroom for adaptation. If there is a civilian aircraft industry, it can more easily retool to make military craft. If the world's biggest vaccine manufacturer is in our backyard, it can license a good vaccine from abroad and mass produce it. Tractor factories can become tank factories. For a contrast, a country like Saudi Arabia or Russia has simple structures of production, and the price system has inferior raw material to work with.

A successful software tool
is one that was used to do something
undreamed of by its author.

-- S. C. Johnson

The most important ingredient is the human capital: the managers, the chemists who know multiple routes to get to a given molecule, the creative people who can hack a machine or a software system to do things that were `undreamed of by its author'. Resilience comes from deep pools of these individuals, who are sparked into self-interested action by the price system. It is equally about the raw STEM knowledge, and about the creative thinking of the business folk who see profit opportunities, who imagine new kinds of deals, who innovate. These pools of capacity lie in the private sector. Even when a government-controlled system has the creative people, it does not have the incentives for them to think, take risks, innovate, and solve problems.

Of particular importance are adjacent products and dual-use technologies. A factory that makes vaccines can be the right starting point to rapidly get a factory to make Covid vaccines. Cylindrical engineering products made using special alloys, for civilian applications, can be rapidly retooled to make ammo. The lowest costs for augmenting supply come from the presence of these neighbours to the desired product.

In normal times, the optimal structure of production tends to become monolithic. The market tends to collapse into a small set of firms and techniques of production. Monolithic methods of production are inherently risky. When crises come along, we see the value of more diversified and more eclectic methods of production. Price surges, in a crisis, create profit opportunities for obscure strategies for production, and obscure producers. These occasional bouts of profiteering serve to keep these obscure firms, these option sellers, alive.

Private sector confidence

The private sector will stand ready with option-like capabilities, it will be alert, it will move mountains to produce when there is a price surge, all in search of one outcome: high profit rates in those brief extreme moments. A society that views supernormal profits as unjust, and tries to expropriate these private firms, is a society where private firms will layer risk premia on top of their ordinary market-based responses. In other words, we would require an even bigger price surge to elicit the supply side response when the probability of expropriation of the firms goes up.

Globalisation

High domestic prices incite imports; the productive capacity of the whole world is brought to bear upon the shortage within one country. Covid vaccine manufacturing in India was about an Indian facility that licensed a British vaccine design, and used numerous imported materials. A deep engagement with globalisation also increases resilience by fostering higher human capital of the elite. An inward oriented economy, with barriers to cross-border activities in the laws and in the minds, is likely to be less resilient.

Storage

If more ammo, oxygen or PPE are held in storage, this creates greater resilience. There is no free lunch; this storage has costs in terms of the opportunity cost of capital, the cost of storage and depreciation. Someone has to pay for this.

Production capacity that has an upside

Consider a factory that makes ammo. If the private person has a contract where there are the assembly lines and staff running at 1 shift, but are ready to jump up to 3 shifts, then there is headroom for a 3$\times$ increase in output. Sometimes assembly lines can be designed in a way where additional workers can be added and the line then runs faster. This can potentially create space for another 2$\times$ increase in output.

In the case of oxygen, firms in the field of industrial gases can have additional equipment on standby, through which medical oxygen output can go up on demand.

As with storage, there is no free lunch. The private firm would have to have contracts with skilled workers in order to be able to surge the production on demand, and design a production system with this kind of headroom. As with the `disaster recovery systems' in the world of software, the principal should randomly trigger these provisions every once in a while, and verify that each agent is indeed able to surge output as promised under contract.

Capabilities in government contracting

When there was a sudden requirement to fly students from Warsaw to India, the best pathway lies in the government rapidly running an auction, where global airlines compete to deliver the lowest price. Surge capacity for the state lies in the combination of (a) A capable and innovative private sector and (b) A state that is able to enter into contracts with private persons.

Is there a role for public policy here?

Fighting wars is a service that is produced by the government. The strategic planners in the field come up with a requirements document such as `We need to be able to sustain a war for 3 months where we are using 100 tubes a day'. Establishing this level of surge capacity is required as part of production of the public good of defence.

In the case of health, what is required is a careful counting of deaths owing to Covid-19, and assessing the number of deaths which are attributable to the shortage of medical oxygen. A careful analysis is then required, where the statistical value of a life is compared against the costs to society of higher surge capacity for oxygen. If a certain enhanced surge capacity for oxygen is able to save lives, while spending less than the statistical value of a life, there is market failure, and then there is a case for public policy to think about state action.

The fact that there is a surge in oxygen demand does not necessarily imply that there is market failure. We can envision private hospitals propositioning health insurance companies and to individuals, saying that they have established the following kinds of surge capacity for oxygen. This is not unlike the work that private hospitals do, in order to assure themselves of electricity in the event of a disaster. We should skeptically evaluate whether we want a government to do something.

Consider the field of masks and other personal protective equipment (PPE) at the early stages of the pandemic. When demand went up by 50 to 100 times, prices skyrocketed. Some policy makers were red in the face and barged into the economy, with export bans, with efforts to supplant the managers of private firms and organise production. But the right response was to do precisely nothing. High prices created near-magical responses by the private sector; there was a surge of import and production, and competition drove down prices.

State intervention that harms surge capacity

When the price system gets going, solving the mismatch between supply and demand through high prices, we often get many calls for state intervention into the working of the economy with tools like price limits and ordering private firms to operate in certain ways. It is ironic that the very feature that incites more production and reduces the demand -- high prices -- is what irritates some people.

Firms will earn supernormal profits in a surge. These supernormal profits are the fair return for (a) The hard work to modify production capacity in a short time; (b) The alertness and risk taking when faced with an incipient surge; and (c) The long years of holding option-like capabilities which are not earning high returns in normal times. When a society begrudges these supernormal profits, and uses state power to expropriate firms, the response of firms is to be less alert, take less risk, do less hard work in modifying production capacity and hold less real options, all of which worsens the problem faced by society in responding to the surge.

To commandeer resources, to order private firms, without proper compensation, is expropriation. During the second wave, many private firms were forced to stop their production in order to transfer oxygen to medical applications. If they were not compensated for their lost production, this constitutes expropriation.

Price controls hamper the very process of healing. High prices kick off the modification of the resource allocation in order to produce more and consume less. When policy makers use state coercion to force transactions to take place at artificially low prices, this reduces both responses. The one thing worse than a price that moves rapidly by a lot is one that does not.

There are always eclectic and opportunistic firms that jump into the fray and reap huge profits when a certain situation presents itself. These firms might even earn nothing at normal times, and just provide options to society. When the state interferes with the 'profiteering', their viability is adversely impacted.

State intervention that gets surge capacity at an excessive cost

One path to having the requisite amount of peak ammunition is to build a large number of public sector factories, which are idle in normal times, where the full cost of a factory is paid and the workers do nothing. While this does get the job done, it is an inefficient path; it does not harness the cleverness of private firms to get the same surge capacity at a lower price.

When there is a shortage in the country, it is tempting to ban exports. This appears to augment supply in the country, and bring down prices, in the short term. But it harms the trust of all firms to produce in India and thus harms India's long-term growth.

The Indian state attacked firms who were importing oxygen concentrators at the time of their peak demand [example]. This amplified the required rate of return for doing this important work.

The most damaging state interventions are those that directly control the resource allocation (e.g. forcing factories to close down so as to grab their oxygen), and in violating the rule of law with outright threats to coerce private persons. When the state becomes such a bull in the china shop, it tends to disrupt the complexity and sophistication of the resource allocation of the market economy. This encourages private people to produce less in India.

The discussion here, of unwise state intervention, is related to the problem of supply chain resilience when faced with Chinese exports of APIs to the Indian drugs industry. There also, it is possible to do clumsy things. Bambawale et. al. 2021 show how to do this better, how to go with the grain of the price system.

Going with the grain of the price system for surge capacity in ammunition

In the field of defence, strategic thinking should ideally generate a requirements document such as `We need to be able to sustain a war for 3 months where we are using 100 tubes a day'. Alongside this, there may be a peacetime requirement of 5 tubes a day, i.e. a peak-to-base ratio of $20\times$. This problem would get handed off to defence economics.

The best way for defence economics to solve this problem is to undertake the following kind of contract:

  1. To ask for multiple private vendors who add up to a peak capacity of 100 tubes/day while actually running every day in peacetime at one-twentieth this rate;
  2. The private firms would find the cost-minimising paths for obtaining this flexibility in production, and they would do this better than a PSU or a government department;
  3. Each private firm would be subjected to random fire drills, where they are asked to suddenly up their production by $20\times$ for a period of $n$ days with $n < 90$.

In this procedure, we have fixed the surge capacity and are procuring on the price. Alternatively, the procurement can fix the price of the tube, and ask for bids which promise the highest surge capacity.

Through this, the energy and intelligence of the private sector would be brought to bear on the problem of obtaining surge capacity for the public goods of defence. It is better to have multiple private vendors, rather than one, so as to avoid single points of failure/attack, and to set off the spiral of quality where private firms compete with each other to deliver bigger surge capacity at a lower price.

This requires complexity in government contracting. Government contracting is a critical homeostatic capability that is required by all states, which works poorly in India. This is an important field for research.

Once contracts are in place, state actors must work within the rule of law: they must not not coerce private persons to behave in ways which were not contracted. Once the Indian state has behaved correctly for a few generations, the private sector will become more comfortable, and will require reduced safety factors in their pricing.



I thank Akshay Jaitly, Amrita Agarwal and Pranay Kotasthane for useful conversations.

Identifying roadblocks in highway contracting: lessons from NHAI litigation

by Charmi Mehta and Susan Thomas.

Introduction

Government contracting is an important foundational process which shapes state capacity. Under Indian conditions, there is limited state capacity in government contracting, which leads to a significant rate of contracting failure. One manifestation of contracting failure is litigation. Defects in the contracting process are often associated with litigation. In addition, litigation leads to delays in project completion, and the prospect of litigation deters some private firms from working with government thus hampering competitiveness in government contracting (Mehta and Uday, 2022). The study of litigation is, therefore, a pathway to creating knowledge in the field of government contracting. An example this was Damle et al, 2021.

In the `contract life cycle approach', government contracting is a pipeline that runs through four phases: I - Contract design, II - Contract award, III - Contract management and IV - Contract closure. We must measure the incidence of defects across the four phases, so as to prioritise resourcing, in the field of policy research, and in the activity of government contracting.

In this article, we construct a novel data-set about contract related litigation at the National Highways Authority of India (NHAI), which is the single agency that does the largest number of government contracts. Using this data-set, we measure the share of each of the phases of the procurement life-cycle in litigation. In our knowledge, this constitutes the first empirical evidence about the role of the phases of the government contracting life cycle in inducing litigation.

Methodology

We hand-construct a data-set which draws upon information from three sources.

Delhi High Court case orders
NHAI contracts and concession agreements give the Delhi High Court exclusive jurisdiction over its contractual disputes from anywhere in India. We find 1165 cases in 2007-2020 that emanate from NHAI contracts. For each case, we observe the date of initiation and disposal, and the names of parties involved.
Extracting facts from court documents
We read 635 out of the 1165 (65 percent) court documents. Using these, we extracted the cause of dispute whenever it was available. These were found to be one of the following: Interim relief, Appointment of arbitrator, Challenging arbitration award, Seeking extension of time, Seeking award enforcement, Restraining NHAI, Reimbursement of costs/Compensation, Challenging claims demanded.
The NHAI Draft Red Herring Prospectus (DRHP)
While filing for the NHAI Infrastructure Investment Trust listing with SEBI, MoRTH was mandated to disclose all large-value disputes involving the NHAI (that are in arbitration), along with a list of all outstanding claims. This had 40 disputes listed, with information about the dates of initiation, cause of disputes, the model of contracting used for the project under dispute, and the names of parties involved. As a self-disclosed data source, this served as a tool to validate observations from the Delhi High Court case orders data.

We find that NHAI accounts for almost 40% of the cases that the Union Government is party to (1165 out of the Union Government's 2912). We also find that cases involving the NHAI have grown at a rate of 17.15% on average during this period. NHAI is the petitioner for about 45% of the cases, while the corresponding figure for other government agencies is about 30%.

NHAI is such a large scale litigant, that reducing the litigation intensity at NHAI would make a different to the working of the judiciary. This work is thus relevant not just for the field of government contracting, but also for the field of the judiciary.

We develop a two-step process for classifying NHAI contract related litigation orders into the four phases of the government contracting life cycle. In the first step, we categorise cases based on causes of dispute as follows:

  1. Related to arbitration proceedings: One set of cases pertain to petitions that (1) seek interim relief from the court under Section 9 of the Arbitration and Conciliation Act 1996, (2) seek the court's intervention in selection of the arbitrator (on account of delays or disagreement in appointment), and (3) appeal awards announced by the arbitral tribunal, which accounted for a significant fraction of the arbitration related matters.
  2. Payments related: This could be about payments, revision of payment terms, price estimation due to changes in scope/ additional components midway into the construction phase or changes in law or policy leading to adverse conditions of work for the private party. Changes in terms of the contract result in parties disagreeing over revised cost estimates or the method of calculating these revisions. In several cases, the delays were related to reimbursements to the private party on account of delays in land acquisition, or operationalising toll plazas. In some cases, the private firm sought restraining orders from the court against the NHAI invoking bank guarantees when there was a delay in the firm achieving financial closure.
  3. Related to wrongful termination/debarment of contractors: Concession agreements embed the terms of contract termination. Prior to the 2020 amendment in the model concession agreement (which allowed a mutual exit clause), the contract perceived termination as default by either side, with the opposite party having a right to compensation and damages. In the process, NHAI often debarred/black-listed contractors from bidding for tenders, a decision that contractors appeal in court.
  4. Reason unclear: For 253 cases, we were unable to classify the cause of dispute since the order was unclear for interpretation. This may reflect blemishes in the Delhi High Court website. We are conservative and classify a case under `Reason unclear' when we are not confident about placing it into the other three categories.

We go on to classify cases into the four phases of the life-cycle of a contract. This mapping is sensitive to the domain; our classification processes are specific to government contracting in highways and infrastructure.

In all the cases that we studied, there was none in the first phase (the "Contract design" phase).

Litigation associated with defects in the pre-award phase involves disputes on the tendering/awarding processes. For example, this includes disputes regarding unfairly awarded tenders, wrongly classifying bids as non-responsive and irregular financial bid-opening processes, all of which are placed within the pre-award phase.

Reading the orders shows that arbitration-related and payments-related causes of disputes take place once the contract has been awarded. This is also true for some of the cases that are classified as wrongful debarment/termination which arises because of the non-performance of contractual obligations. Since these disputes arise after the contract is awarded, we classify these cases in the post-award phase. Some of the orders contain details about the type of contract between the parties. For example, SPVs are often formed for the construction of a highway, and are named '... Tollway Pvt. Ltd.' which indicates a legal entity that was formed for the purposes of executing a contract or an award. This can be identified from the names of the parties, even if the specific causes of dispute might be unclear in some cases. We use this information in classifying cases under the post-award category.

Traditionally, cases pertaining to payment delays are categorised within the post-completion phase. But in the case of NHAI contracts, payment schedules vary across the commonly used models of Hybrid Annuity (HAM), BOT-Annuity models and EPC. Payments are made to the concessionaire either at the start of construction, or at regular frequencies such as semi-annually, with payments usually linked to the achievement of project milestones. We place such cases in the post-award phase rather than in the post-completion phase. Similarly, case orders that seek to resolve arbitration-proceedings related disputes are also categorised as post-award since these disputes arise from a contractual relationship between two parties to approach alternate dispute resolution mechanisms, which can only exist in the event of a contract being awarded.

Findings

Table 1: Number of cases by drivers of litigation

Cause of dispute NHAI as petitioner Firm as petitioner Total
Arbitration proceedings related 123 137 260
Payments related 15 75 90
Wrongful termination/ debarment 1 31 32
Total 139 243 382

In Table 1, we summarise the number of cases falling under Categories 1-3 listed above. We further break up the cases under each category as the number of cases where NHAI was the petitioner and where the private firm was the petitioner. A large volume of cases are a) arbitration-proceedings related, and 2) payments-related (delayed payments). We see that arbitration related matters are nearly 70 percent of the sample of case orders. In this category, both the NHAI and the firm are almost equally likely to petition an arbitration matter. In the other two categories, NHAI is more likely to be litigated against.

The analysis based on case classification into phases of the contract life-cycle is presented in Table 2. This shows the number of disputes which we classify as falling within the pre-award, post-award, and post-completion phase. In this table, we present both the actual number of cases identified in a given phase, as well as the share of the cases as a fraction of the total.

Table 2: Share of disputes by the phase of the contract life-cycle

Phase of the contract life-cycle Share (% of Total) Number of cases
Pre-award 1.75 11
Post-award 66.14 420
Post-completion 0.15 1
Unclear/No data recorded 31.95 203
Total 100.00 635

Table 2 shows that over 65% of the cases are concentrated in the post-award or the contract management phase.

Discussion

In this article, we have analysed one important government contracting organisation, and attributed its litigation into the four phases of the contract life cycle. The results diverge from the commonly held view that the problems in public procurement arise because of problems with Phase 2 or L1 tendering. This analysis suggests that the hot spot in litigation is Phase 3 or the Contract management phase. This suggests that policy researchers and policy practitioners should allocate greater resources on strengthening this phase.

Another finding is that there is a large volume of arbitration-proceedings related disputes, which indicates the limitations of alternate dispute resolution (ADR) mechanisms. Strengthening ADR mechanisms in India would result in less litigation.

Recent reforms of the government contracting process have initiated movement on this front. The General Instructions on Procurement and Project Management (2021) built provisions to address concerns around arbitration and dispute resolution. It acknowledges that the majority of the cases appealed by government agencies end up being awarded in favour of the opposite party. In this context, the General Instructions emphasize the need for the government to actively take steps to minimise litigation. Further, it inserts Rule 227A in the General Financial Rules (GFRs) 2017 for Arbitration Awards, which mandates that the concerned ministry pay 75% of the amount of the arbitral award against the bank guarantee, even if it contests the award in a court of law. This is a step in the right direction as it takes away the incentive for either party to benefit from prolonging a dispute or appealing to a higher authority to delay payment of claims awarded by the arbitral tribunal.

References:

Damle D., Gulati K., Sharma A., and Zaveri-Shah, B. Litigation in public contracts: some estimates from court data., The LEAP Blog, May 2021.

Mehta C., and Uday D., How competitive is bidding in infrastructure public procurement? A study of road and water projects in five Indian states, The LEAP Blog, March 2022.

Charmi Mehta is a researcher at XKDR-Chennai Mathematical Institute and Susan Thomas is professor at the Jindal School of Business and a researcher at XKDR Forum. The authors thank Shailesh Pathak and Bhargavi Zaveri-Shah for their inputs in shaping the study, and Yajat Bansal and Diya Mal for their research assistance in this study.

Sunday, July 03, 2022

Measuring financial inclusion: how much do households participate in the formal financial system?

by Geetika Palta, Mithila A. Sarah and Susan Thomas.

Measuring the impact of financial inclusion

Households use financial instruments and financial markets to achieve their lifetime objectives. These include being able to smooth consumption over time, being able to withstand shocks, and pursue entrepreneurial opportunities to gain income mobility. Financial inclusion refers to such access to finance for a larger subset of the population (e.g. Rao, 2018). Financial policy makers have pursued financial inclusion for many decades. In recent years, the rise of ESG investors has bolstered private sector interest in financial inclusion.

For policy makers, for financial firms, and for ESG investors, there is thus an interest in the measurement of financial inclusion (Sarma M., 2016; UNEP FI, 2021). The field of measurement of financial inclusion is under-developed. While there is high interest in building such measures (RBI, 2020; El-Zoghbi, 2019), there are debates about methods and no single measure has been widely accepted (Nguyen, 2021).

Financial inclusion should improve the life of the household through smoothing consumption, withstanding shocks to income and helping the household achieve income mobility to a higher sustained level of consumption. For example, learnings from a financial literacy program in the Philippines show how Filipino households obtained income mobility (Monsura, 2020). These households learned how to take advantage of the economic opportunities through savings, investment, insurance, and entrepreneurship. Access to formal financial services and the ability to use them enables the households build wealth and generally live a financially secure life.

An inputs-outputs-outcomes framework

The inputs-outputs-outcomes framework is valuable in many aspects of policy thinking. As an example, in a domain like education, the input is school buildings, the output is children spending hours in school, and the outcome is the change in their knowledge (Banerji et al., 2013).

This approach is valuable in the field of financial inclusion also. The input is household participation in formal finance (such as account opening or purchasing health insurance); the output is the intensity of transactions (how frequently the account is used or whether the insurance premium is paid on a regular basis) and the outcome is the impact on economic well-being.

This perspective upon financial inclusion guides measurement methods for financial inclusion. Measurement of financial inclusion needs to measure inputs (presence of various financial products and services in the household portfolio), outputs (the use of financial products in achieving household objectives) and outcomes (stability of consumption and income mobility).

Done right, such measures can facilitate a deeper understanding of the impact of financial inclusion on the economic well-being of a household. These measures can help identify gaps in financial inclusion, both in terms of missing products in the household financial portfolios, as well as excluded household groups. For ESG investors, these measures can play a role in their principal-agent problems with portfolio companies.

In this article, we propose and implement a simple financial inclusion input measure, which is the household participation in the formal financial sector, calculated using the sample of households in the CMIE CPHS data. With this, we show some important facts about financial inclusion inputs in India.

Difficulties of conventional measures

In the early stages of measuring financial inclusion, crude proxies were used for measurement at the level of the economy, such as M2 (cash, demand and time deposits) as a percentage of GDP. Later, more systematic data collection about household holdings of financial assets began (Beck, 2016). Most of these measures were typically country-level aggregates organised around financial service provider (FSP) or one class of financial product (RBI, 2017). While aggregates at the country level are useful, they can mix up usage by some households and absence by others. What would be most useful is to construct financial inclusion measures at the level of a household, pulling together a full picture of the financial activities of the household (Campbell, 2006).

More often than not, there has been a bank-orientation in these measures with focus on number of bank accounts, bank branches, number of ATMs and amount of bank deposits. But there is much more to financial inclusion than banking. Gupta and Sharma (2021) point out that measuring ownership of bank accounts alone tends to overestimate and present an incomplete picture of financial inclusion as it neglects access to and use of the full range of financial products. Over time, the focus of financial inclusion has shifted towards a larger set of financial assets and usage of digital payment systems (RBI, 2020).

The construction of financial inclusion measures at the level of a household pre-require a capture of such information from households themselves. There are a few rare instances where countries have administrative data from which asset portfolio by households can be constructed (Calvet et al., 2007; Andersen et al., 2020). Most countries do not have such data on household portfolio of financial instruments (Badarinza et al., 2016; IFC, 2011). Over the last decade or so, household surveys have emerged that record household portfolio of financial instruments. Most of these have been one time surveys or surveys done at low frequencies. For example, in India, the NSSO AIDIS captures household level participation in financial systems once in 10 years.

Constructing a household `Financial Participation Score' (FPS) using CPHS

An important household survey that is conducted thrice a year over a sample of 170,000 households is the Consumer Pyramids Household Survey (CPHS), by the Centre for Monitoring Indian Economy. Given India's high economic growth rate and the rapid pace of change in the last few decades in finance, this survey makes possible new insights into financial inclusion of Indian households in a timely and geographically dis-aggregated manner.

The CPHS has member-wise characteristics and household characteristics such as income and expenditure of households, what assets they own and whether they have borrowings. Household data on financial assets owned comes from the ''People of India database'' and the ``Household Aspirational India database'' in CPHS. In the former, households are asked questions on ownership (Yes/No) of four different financial instruments, while the latter measures outstanding investment (Yes/No) in six financial instruments. We use the following variables to measure the financial participation of a household:

  • Household ownership of at least one bank account (Bank), at least one health insurance (HI), at least one life insurance (LI), at least one employee provident fund account (EPF).
    This captures four components of financial inclusion.
  • Outstanding investment at a household level in fixed deposit (FD), Kisan Vikas Patra (KVP), National Savings Certificate (NSC), Post Office Savings account (POS), Mutual Funds (MF) and Listed Shares (LS).
    This captures six components of financial inclusion.

Put together, there is data about 10 financial instruments -- all zero/one values -- that households hold at a point in time. We define a Financial Participation Score as sum of the values divided by 10. This gives the household an FPS that runs from 0 to 1. For example, an FPS value of 0.3 indicates that the household owns three of the ten financial instruments.

The CPHS data on household holding of the 10 financial instruments is captured three times a year in three ``waves'' where each wave is completed over four months and surveys about 170,000 households. In each year, Wave 1 consists of January, February, March, April 2021; Wave 2 has May, June, July, August and Wave 3 has September, October, November and December. Households are generally measured in a consistent month slot within each wave thus generating a regular cadence in the time-series for each household.

All the 10 instruments used in this calculation involve households carrying consumption from the present into the future. In this article, we do not include debt-related variables in calculating financial participation, even though borrowing is one form of finance used by many households. For one, debt is multi-dimensional. It can be from different sources (formal vs. informal), have different maturities, be driven by different purposes. While all debt involves carrying consumption from the future to the present, the impact of debt on the future well-being of the household can vary. Some debt is for short-term consumption smoothing, possibly at the cost of lower consumption in the future. Other types of debt may lead to higher income in the future if they are used to build enterprise. Given this multi-faceted nature of household debt, it's inclusion is left for downstream research.

We construct an unbalanced panel data-set of household FPS at the wave level, for 2014-2021, with three waves per year. The number of households observed varies from 76,386 (during the lock-down in 2020) to 1,49,160 (2018). The CPHS is a stratified random sample. However, for the purpose of this first exploration of basic facts, we have reported unweighted summary statistics.

Some basic facts about the FPS

The household FPS is calculated for each wave. The annual FPS of a household is calculated as the maximum value of FPS observed for the household across all the waves for which it was observed. The summary statistics of annual household FPS values are presented in Table 1 for each year of the panel data-set.

Table 1: Summary statistics of household FPS, from 2014 to 2021


2014 2015 2016 2017 2018 2019 2020 2021
Min 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
25th 0.1 0.2 0.2 0.2 0.2 0.2 0.2 0.1
50th 0.2 0.3 0.3 0.3 0.3 0.3 0.3 0.2
75th 0.3 0.3 0.4 0.3 0.4 0.4 0.4 0.3
Max 0.9 0.9 1.0 1.0 1.0 1.0 1.0 0.9

For most of the years, 50 percent of the households hold 3 or fewer instruments. This holds steady for the 6 year period, for most part. There continue to be households with FPS of 0. This implies that there continue to be households that do not even have bank accounts in this sample.

There have been minor shifts in financial participation of the households in this period. The COVID-19 pandemic lock down of April to June 2020 appears to have an adverse impact. By 2021, the median household has dropped from holding 3 instruments to 2. This is a consistent drop -- the 25th percentile household have dropped from 2 to 1 instrument, and the 75th percentile household has gone down from 4 to 3.

We next examine the cross-sectional variation in household participation. For this, we categorise all households into five groups: those with (1) FPS less than 0.2, (2) equal to 0.2, (3) equal to 0.3, (4) equal to 0.4, and (5) greater than 0.4. Figure 1 shows the fraction of households in each of these FPS categories, in each wave.

Figure 1: Distribution of households by categories of FPS

Figure 1 shows that there was an increase in household financial participation in the early part of this sample, from 2014 up until the end of 2017. (The areas under the sum of FPS categories >= 2 have dropped in this period.) In 2018 and 2019, there was no change in the fraction of households across the defined categories. The changes of 2014-2016 appear to reverse from the second half of 2020 onwards. By 2021, the fraction of households with FPS >= 0.3 is nearly the same as the values seen in 2019.

What was happening at the level of the individual instruments?

In Figure 2, we go below the aggregate FPS into portfolio of individual instruments, including bank accounts, fixed deposits, pensions, post office savings, health insurance, life insurance, mutual funds and listed shares. (We do not include the household holdings of KVP and NSC because these fractions were very small compared to the selected eight instruments in the figure.)

Figure 2: Distribution of households portfolio of individual financial instruments by wave (log scale)

Health insurance had the highest growth (10 percent of households holding to 40 percent of households holding in the sample in a wave). At the same time, life insurance saw a drop (from 60 percent of households holding to 40 percent of households in the sample holding this in a wave). Post office savings saw an increase (from 8.5 percent of household holding to nearly 20 percent of households holding) while pensions saw a decrease (from 25 percent of household holding to around 18 percent). While the numerical values are small, there was strong growth in mutual funds and listed shares.

How different is financial inclusion for rural vs. urban households?

How does the financial participation of urban households compare to rural households? In the following Figure 3, we examine the distribution of rural and urban households in the four FPS categories presented in Figure 1.

Figure 3: Distribution of rural and urban households by categories of FPS

The figures show that the distribution of rural households tend to have lower financial inclusion compared to the urban households. More interesting is the difference in the evolution of financial inclusion between these two groups. Both rural and urban households saw increasing financial participation in 2015 and 2016 compared to 2014. However, financial participation of rural households stalled at the end of 2016, while urban households contend to grow their financial participation. Financial participation for both rural and urban households worsened first in 2018, and then more sharply in 2020, at the time of the pandemic.

We also examine what are the differences in financial instruments holdings behind the variation that we see in the financial participation of rural and urban households. From Figure 4, we can see that rural and urban households are similar in their holding of bank accounts, fixed deposits and post office savings. But they are distinctly different in their holding of EPF, mutual funds and listed shares, where there is a higher fraction of urban households holding these instruments compared to rural households.

Figure 4: Distribution of rural and urban households' portfolio of individual financial instruments

Figure 4 also shows us that the growth in fraction of households holding individual instruments vary between rural and urban households. There was a higher growth in fraction of rural households holding health insurance (from 5 to 40 percent), while for urban households this was lower (from 10 percent to 40 percent). There was a drop in the fraction of rural households holding life insurance compared to no change in the fraction of urban households holding these.

This tells us two pertinent aspects of the growth of financial participation across rural and urban households: first, financial participation by rural households appear more vulnerable to external shocks -- such as demonetisation, the ILFS-NBFC crisis and the pandemic -- than urban households. Second, there is some variation in what types of instruments rural households tend to hold compared with urban households.

In the CPHS sampling strategy, there is a roughly two-times over-weighting of urban locations. The simple summary statistics shown in this article (i.e. unweighted estimates) are problematic; for more precise estimates all summary statistics require appropriate weighting. It is hence particularly useful to see the urban and rural values separately, as has been done here.

Conclusions

It is widely believed that improvements in financial inclusion will translate into reductions of consumption volatility and increased odds of improved lives. Greater research is required on measuring the strength of these relationships. In the standard recipe of phenomenological research, we require measurement of a phenomenon, and then it becomes possible to analyse the causes and consequences.

An important missing link in the field of financial inclusion are tools for measurement. In this article, we have shown a first and simplest measure, an input measure, about use of the formal financial system by households. This measure can be computed at the household level, three times a year, in the CMIE CPHS survey database.

In the summary statistics shown here, there have been only small changes in the overall average FPS over the years under examination. The median value for urban households was 0.3 and the median value for rural households was 0.2. We see a visible decline of the FPS in the lockdowns of 2020, and in the post-pandemic economic recovery, the FPS has come back to near pre-pandemic values. These results suggest numerous questions about causes and consequences, which need to be explored in downstream research.

This ability to observe the FPS at the level of a household enables new kinds of academic research, new kinds of feedback loops for policy makers, and definitions and measurement to help ESG investors overcome principal-agent problems between the investor and the fund, and the fund and the portfolio company.

References:

Andersen, S., Campbell, J. Y., Nielsen, K. M., & Ramadorai, T. (2020). Sources of inaction in household finance: Evidence from the Danish mortgage market, American Economic Review, 110(10), 3184-3230.

Badarinza, C., Campbell, J. Y., & Ramadorai, T. (2016). International comparative household finance, Annual Review of Economics, 8, 111-144.

Banerji, R., Bhattacharjea, S., & Wadhwa, W. (2013). The annual status of education report (ASER), Research in Comparative and International Education, 8(3), 387-396.

Beck, T. (2016). Financial Inclusion–Measuring progress and progress in measuring.

Calvet, L. E., Campbell, J. Y., & Sodini, P. (2007). Down or out: Assessing the welfare costs of household investment mistakes, Journal of Political Economy, 115(5), 707-747.

Campbell, J. Y. (2006). Household finance, The Journal of Finance, 61(4), 1553-1604.

El-Zoghbi, M. (2019). Toward a New Impact Narrative for Financial Inclusion, CGAP, 2019.

Gupta, S., & Sharma, M. (2021). A Demand-Side Approach to Measuring Financial Inclusion: Going Beyond Bank Account Ownership, Dvara Research Working Paper Series No. WP-2021-05.

Monsura, M. P. (2020). The importance of financial literacy: Household's income mobility measurement and decomposition approach., The Journal of Asian Finance, Economics and Business, 7(12), 647-655.

Nguyen, T. T. H. (2021). Measuring financial inclusion: a composite FI index for the developing countries , Journal of Economics and Development, Volume 23, Number 1, pp. 77-99, 2021.

Nilekani N (2019). Report of the High Level Committee on Deepening of Digital Payments, Reserve Bank of India.

Rao, K. S. (2018). Financial inclusion in India: Progress and prospects , The Ideas for India blog, 2018.

RBI (2017). Report of the Household Finance Committee on Indian Household Finance, 24 August 2017.

RBI (2020). National Strategy for Financial Inclusion 2019-2024 .

Sarma, M. (2008). Index of financial inclusion Working paper No. 215, ICRIER, New Delhi.

Sarma, M. (2016), Measuring financial inclusion using multidimensional data, World Economics, 1 Ivory Swuare, Plantation Wharf, London, UK, SW11 3UE, vol. 17(1), pages 15-40, January 2016.

International Finance Corporation (IFC) (2011). Financial inclusion data: assessing the landscape and country-level target approaches. The World Bank, 2011.

United Nations Environment Program Finance Initiative (UN EPFI) (2021). 28 Banks collectively accelerate action on universal financial inclusion and health, 2 December 2021



Geetika Palta, Mithila Sarah and Susan Thomas are researchers at the XKDR Forum. We thank Ajay Shah and three anonymous referees for valuable comments and suggestions.

Wednesday, June 15, 2022

Reconsidering SEBI disgorgement

by Renuka Sane and S. Vivek.

SEBI disgorgement is a regulatory remedy to recover wrongful gains from entities that have violated securities laws. It is justified based on the equitable principle that no one should benefit from their own wrong. This seems like a non-controversial, even obvious, ground for regulatory action that has 'compelling intuitive appeal'. However, there are basic conceptual issues that are not clearly settled, not just India but in other jurisdictions as well. For example, the U.S. Supreme Court has considered three cases on disgorgement over the last few years - in one case, it held that disgorgement was beyond the powers of the Federal Trade Commission, overruling decades worth of practice, and in another, upheld Securities and Exchange Commission's power to seek disgorgement but with important conceptual restrictions.

These are trends that the Securities and Exchanges Board in India (SEBI) should be watching carefully. Lack of conceptual clarity about the remedy can put years of regulatory action at risk, if the basis of the remedy is questioned in a case before superior Courts. Further, a study of how SEBI orders are interpreting disgorgement powers and if they are consistent with the conceptual justifications is critical. SEBI disgorgement does not have any statutory limit - the order can direct recovery of all the wrongful gain, whatever they may be. This exercise of of vast discretion by SEBI Whole-Time Members (who are executive members of the regulator and typically do not have substantial judicial training), without transparent statutory or conceptual guidance, raises regulatory governance concerns.

In a new working paper, Reconsidering SEBI disgorgement, we study disgorgement from three perspectives:

  1. The theory of disgorgement: Disgorgement is a distinct remedy that must be distinguished from other remedies such as compensation, restitution, and penalties. Disgorgement is different from compensation because compensation is focussed on the loss suffered by claimants whereas disgorgement is focussed on the gains made by the wrongdoer. While disgorgement and restitution are both gain-based remedies, there is a subtle yet important difference. Restitution is focused on reversing a wrongful gain of the defendant based, for example, on a wrong or incorrect transfer from the plaintiff. Here, the (wrongful) gain made by the defendant is equal to the loss suffered by the plaintiff - the property in question (money, for example), is returned to status quo. Disgorgement, on the other hand, strips the defendant of its gains, even if such gains are not made from the plaintiff, and even if the plaintiff does not suffer any loss. Accordingly, the loss suffered by the plaintiff need not correlate to the defendant's gain that is clawed back through disgorgement. The objective for disgorgement is to have a deterrence effect, and not to merely reverse an illegal transfer. Penalties are also generally imposed for the purpose of deterrence, among others. However, while disgorgement amounts must be equal to the gains made by the wrongdoer, penalties can be imposed merely on the basis of the violation and need not correlate exactly to the gains, if any, made by the wrongdoer.

  2. The evolution of disgorgement at SEBI: SEBI had even in the early years tried exercising powers to claw back illegal gains (disgorgement), or compensate victims in insider trading cases, with mixed success at appellate fora. Subsequently, parliamentary and expert committees over the years have recommended providing SEBI with clear powers to trace the illegal gains made by wrongdoers and return such gains to their victims. Before such powers could be formalised through statute, disgorgement was used by SEBI Whole-Time Members (WTMs) as a quasi-judicial innovation in their orders,and received approval from the Securities Appellate Tribunal. Since, at that time, SEBI WTMs did not have the power to impose monetary penalties, SEBI disgorgement was justified as a 'remedial' power which only returns the wrongdoer to status quo, and hence can be distinguished from a punishment. Further, as there was no express statutory provision at the time for SEBI disgorgement, it was traced back to 'equitable' powers of SEBI WTMs.

    In 2014, the Securities and Exchange Board of India Act, 1992 (SEBI Act), was amended to clarify that SEBI disgorgement was part of SEBI's remedial powers. The amendment also stated that the amounts so clawed back are not to be deposited with the Consolidated Fund of India as in the case of penalties; instead, they are retained by SEBI's Investor Protection and Education Fund, to be used in terms of SEBI's own regulations. Interestingly, despite tracing its origins to Parliamentary and expert committees which discussed disgorgement powers in the context of using the proceeds to compensate victims, the amendment did not require SEBI to even attempt to distribute the amounts to victims. Since then, SEBI's power to direct disgorgement without clear statutory limits has been entrenched. Gradually, SEBI also received judicial recognition for its power to impose interest on the disgorgement amount. These rates are calculated from the date of the violation, sometimes going back 10 years or more (as opposed interest on penalties which is typically calculated from the date of non-payment after the SEBI order). Further, the initiation of proceedings for disgorgement or penalties, remains with SEBI and it is unclear how it is exercised.

    These vast powers are conferred on the regulator on the basis that SEBI disgorgement is only 'remedial' and is returning the wrongdoer to status quo. The use of the term 'disgorgement' while at the same time emphasising the return to status quo creates some confusion between the related albeit distinct remedies of disgorgement and restitution. In this context, we study whether in practice what kind of remedy SEBI disgorgement actually is, regardless of its nomenclature. Further, as a legal matter, returning the wrongdoer status quo is critical as a point of distinction from SEBI penalties; if the wrongdoer is left worse-off, it could be argued that SEBI disgorgement is a penalty by another name.

  3. The practice of disgorgement at SEBI : If SEBI's case for disgorgement is based on clawing back illegal gains and returning the wrong-doer to status-quo, do they actually do so? We use all the SEBI disgorgement orders between January 1, 2018, and July 15, 2021, and find that in 9% of the cases there is no finding that the noticee has made a benefit or avoided a loss, and yet noticees have been ordered to disgorge. In none of the cases is there a finding that the direction brings the noticee back to status-quo and does not leave them worse off - a critical element in the justification for SEBI disgorgement and its characterization as a remedial power. Further, it is interesting that SEBI disgorgement is usually used for insider trading, and fraudulent trading offences, for which the SEBI Act allows penalties to be issued up to three times the profits made. Why is disgorgement, and not penalties, being used in these cases?

Our results suggest that lawmakers and the SEBI Board must review how SEBI disgorgement is conceptualised and what goals it serves. It should scrutinise how disgorgement orders are being issued under the existing framework so that they are consistent with the justifications for remedial measures (such as, allowing deductions for legitimate expenses and a transparent and careful system to determine causation of the gains from the wrong). A holistic look at remedies available for securities law violations is required so that they serve all the goals required for stakeholders - deterrence, compensation, and restitution.


S. Vivek is a researcher with the Regulatory Governance Project at the National Law School of India University, Bengaluru. Renuka Sane is a researcher at NIPFP. Author names are in alphabetical order.

Friday, June 10, 2022

Threats to legal certainty in government contracting by electricity distribution companies

by Akshay Jaitly and Ajay Shah.

A battle in Andhra Pradesh, 2018-2022

In September 2018, electricity distribution companies in Andhra Pradesh (Discoms) filed a petition before the Andhra Pradesh Electricity Regulatory Commission (APERC) to reduce the feed-in tariff for wind power projects (that had been determined under Section 62 of the Electricity Act, where the regulator sets the price). Another petition was filed requesting APERC to revise the tariff payable by Discoms under solar power PPAs (this time discovered under Section 63 of the Electricity Act, where there is a competitive bidding process). The argument made by Discoms was that the tariff discovered in other states pursuant to competitive bidding was lower than the tariffs statutorily determined in Andhra Pradesh. There are also newspaper reports about the state load dispatch centre (SLDC) curtailing output by renewables generators, ostensibly in the interests of grid safety.

These attempts at reneging on contracts, by the state, go against basic notions of sanctity of contracting and legal certainty. When X contracts with Y, both are bound by and obliged to fulfil the terms of the contract, regardless of future fluctuations of prices and technology. The Indian Contract Act, 1872, and a line of case law under it, gives no space for either X or Y, as private persons, to renege on a contract because better prices had come about somewhere else in the economy.

It is also settled law that when the state enters into a contract, it does so in a commercial capacity and not as the sovereign. If the Indian state purports to renege on contracts in this fashion, it deepens the problems of the state as an untrusted counterparty, and fewer private persons will be willing to do business with the state in the future. This would harm the prices at which the state is able to enter into contracts.

As an example, consider a Jan 2022 story by Kailash Babar in the Economic Times, about NHAI terminating a contract with IL&FS which had been established in 2013. NHAI did not just walk away: it paid IL&FS Rs.891 crore for the privilege of terminating the contract. Concessions typically have a formula for termination compensation in three scenarios – authority default, no fault and concessionaire default. These are expressed as percentages of debt due plus some equity return and some other terms.

Attempts at reneging on PPAs elsewhere in India

This experience from Andhra Pradesh is actually not unique. Discoms and regulators in Karnataka, Uttar Pradesh, Jharkhand and Tamil Nadu have subsequently attempted unilateral termination or renegotiation of renewable energy tariffs under validly executed PPAs.

Punjab took this one step further in November last year, by introducing and unanimously passing legislation to get out of its PPA obligations. A few months ago, we wrote about the Punjab Renewable Energy Security, Reform, Termination and Re-Determination of Power Tariff Bill passed in the Punjab Legislative Assembly. This law seeks to renege on PPAs that the Punjab state Discom had voluntarily entered into, on the basis that these created too heavy a financial burden on the state. This attempt by the state, to have immunity from contract performance, is under legal challenge.

A seminal skirmish took place a while ago, in Gujarat in 2013, where the Appellate Tribunal for Electricity had held with reference to the actions of a Discom in Gujarat, that a PPA could only be reopened for "giving thrust to renewable energy projects and not for curtailing the incentives". In other words, PPAs could not be reopened to reduce tariffs. In this case, Gujarat Urja Vikas Nigam Ltd (GUVNL) had filed a petition before the Gujarat Electricity Regulatory Commission (GERC) in 2013, asking for a revision in solar tariffs determined by the commission in its 2010 order on the grounds of reduced customs and excise duties, which would justify a downward revision of the tariffs. The GERC dismissed GUVNL’s petition and its decision was upheld by the Appellate Tribunal for Electricity (APTEL) on appeal. APTEL held that since GUVNL had not established that there is a legal right available to it to seek a redetermination of the tariff by reopening the PPA, the GERC would not be expected to revisit the generic tariff ‘to dis-incentivise the project developers and consequently discourage future investment in the sector’.

How the Andhra Pradesh story played out

Despite this, solar and wind power developers challenged this in the High Court of Andhra Pradesh. A single judge bench of the High Court dismissed the Discoms' petitions in September 2019, with a direction to APERC to decide the issues raised by the developers.

But the High Court directed the Discoms to pay an interim tariff (lower than the tariffs under the PPAs) until APERC adjudicated the matter. The legal foundations through which the court chose to go against contract law are not clear. This created tremendous commercial difficulties in the industry. In some instances, there are reports that even this lower interim tariff was not being paid by the Discoms, causing further distress to power generators.

The typical renewables project is a tight arrangement of capital and PPA, with little room for contracting wobbles. Once the predictability of cash flows was disrupted, some of the generating assets were classified as 'non-performing'. The generators tried to go to court to force lenders to not do so.

This problem then showed up at a Division Bench at the Andhra Pradesh High Court. The case played out over three years. The Division Bench held that:

  1. The tariff under concluded PPAs cannot be re-negotiated;
  2. Financial difficulty of Discoms is not a ground to permit non-performance of the PPAs or to reduce the tariff set out under the PPAs;
  3. A tariff determined through competitive bidding process under Section 63 of the Electricity Act cannot be re-determined; and
  4. Since renewable energy plants operate on a ‘must-run basis’, any arbitrary curtailment of power by the state load despatch centre without notice and not based on grid security or safety reasons is illegal.

This was a salutory reaffirmation of the foundations of commercial law: Contracts must be honoured, statutory processes cannot be unilaterally set aside, power validly contracted under a PPA can only be curtailed for technical reasons. At the same time, in a well functioning market economy, these events from 2018 to 2022 -- and the associated commercial consequences for private persons -- should have never taken place. Every investor looks at this fracas and chooses a somewhat higher risk premium for doing business in India.

Implications for the Indian legal system

It is important to analyse what shapes these attempts at state immunity from contract law. In what ways can laws be amended, or principles be evolved, so that such attempts are eliminated or at least minimised?

Perhaps the Andhra Pradesh Discoms will appeal to the Supreme Court of India in this matter; perhaps the challenge to the Punjab PPA law will find its way to the Supreme Court. It is then interesting to envision: What is the Supreme Court order that can usefully underline the foundations of the extant contract law, and thus reduce the incentives to embark on such manoeuvres?

While a Supreme Court order in this regard might act as a deterrent in the future, the problem lies in the culture of government institutions, who are conditioned to exhaust all available means of reducing costs, irrespective of the merits of their position and the chances of success, out of fear of vigilance authorities. A solution would be for the government to develop guidelines and instructions setting out the bases on which appeals should be pursued or not.

In developing such guidelines, some of the questions for the state to consider would be as follows: Suppose the probability of success of such attempts at renegotiating are 0. Is it still efficient for a state government to initiate it? As Karan Gulati and Shubho Roy emphasise in a forthcoming paper, could it be that the time value of money that is used in Indian court orders make it efficient for the state government to embark on litigation that it has no possibility of winning? We need to also analyse the Indian justice system from this point of view, and identify the reforms through which the incentive of a state government is reshaped.

Implications for electricity policy

As we have argued before, the Indian electricity sector has suffered from difficulties for a long time, but the recent years represent an escalation of stress to a different level. This comes from the combination of low price renewables, volatility in fuel costs, the impact of ESG investors abroad upon electricity purchase by large Indian firms, and the accelerating exit of commercial and industrial buyers from discoms.

It is sometimes comforting to think that discoms in India have always had problems. But the problems seen today are worse. Faced with extreme stress, there is an appetite for extreme measures. When the policy process is weak, there is a greater likelihood of poorly designed policy measures being adopted, such as attempts at reneging on contracts. When even a few discoms engage in such behaviour, this reduces the investability of the Indian electricity sector in areas that have connections with the Indian state.

We should see each of these eruptions as illustrations of the underlying stress, and reorient ourselves towards the required fundamental electricity reform.


We thank Charmi Mehta for research assistance on this article.