## Thursday, January 19, 2023

### Examining grievances and redress for pension products

by Vimal Balasubramaniam, Aishwarya Gawali, Nancy Gupta, Renuka Sane and Srishti Sharma.

In a previous article, Examining grievances and redress for banking products, we studied the nature and extent of grievances for banking and payment products in India. We also evaluated whether grievance redress mechanisms worked, and what impact grievances had on the usage of products. In this article we study similar questions for contributory pension products. The analysis is based on a survey of 21,355 respondents that we conducted in five states including Maharashtra, Bihar, Haryana, Madhya Pradesh and Andhra Pradesh.

### Measuring grievances

In the survey, we first ask if the respondent is using or has ever used contributory pension products. These would typically include the General Provident Fund (GPF), the Public Provident Fund (PPF), Employee Provident Fund (EPF) and the New Pension System (NPS). The study explicitly excluded various defined benefit pension plans, such as old age pensions, widow pensions, and disability pensions. We then ask the following questions:

1. If the respondent had faced an issue with any of the contributory pension products in the last 12 months?
2. If yes, what was the latest/most recent issue with the pension product?
3. Did the respondent complain after encountering the grievance?
4. Was the issue resolved after the first complaint?
5. If no, was the complaint escalated further?
6. If the complaint was escalated, was the issue resolved upon escalation?
7. If the respondent did not complain, what was the reason for not complaining?
8. Finally, what was the impact of the grievance on their usage of pension products?

We only consider complaints registered with the financial service provider or pension regulators. We do not include complaints filed in the police station or consumer courts as this is not in the ambit of the regulatory grievance redress system. Our questions do not pertain to any specific pension product. The results, therefore, are a reflection of the overall system of grievance redress, and not of any particular scheme.

Before we describe our results, it is useful to present the existing grievance redress mechanisms in the pension ecosystem. It is also important to note that while certain pension regulatory bodies report the incidence and resolution of grievance, there is no official consolidated statistic on the number of grievances for contributory pension schemes in India. Table 1 provides a snapshot of the governing regulatory bodies and grievance redress mechanisms (GRMs) of some of the major pension products that are relevant to our study.

Table 1: Pension products, regulatory body and GRMs
Pension Product Regulatory Body Grievance Redress Mechanisms (GRMs)
General Provident Fund (GPF) Department of Pension and Pensioner's Welfare under the Ministry of Personnel, Public Grievances and Pensions Online Grievance Lodging and Monitoring System at the Office of the Comptroller and Auditor General of India
Public Provident Fund (PPF) Department of Post of India Centralised Public Grievance Redress and Monitoring System (CPGRAMS) along with a dedicated Grievance Handling Cell accessible via call and email
Employee Provident Fund (EPF) Employees' Provident Fund Organisation (EPFO) EPF i-Grievance Management System (EPFiGMS)
New Pension System (NPS) and annuity schemes Pension Fund Regulatory and Development Authority (PFRDA) A multi-leveled Grievance Redressal System

Our analysis thus pertains to the use of contributory plans, which include, but are not limited to the schemes mentioned above.

### Results

Our sample comprises of 21,355 respondents. 622 (2.92%)individuals reported having used a pension product. This is not surprising given that coverage through mandatory occupational pensions is low. However, there appears to be growing demand for micropension among poor families. This is reflected in our sample as well. Respondents with annual family income of less than one lakh rupees formed the largest share of pension users. 50% (314 out of 622) of pension users had an annual family income of less than one lakh rupees. 30% (187 out of 622) of the pension users had an annual family income between 1 to 3 lakh rupees. The remainder 20% had an annual family income of more than 3 lakh rupees.

### Extent and nature of grievances

Of the pension users, about 11.4% (71 out of 622) reported having faced grievances related to pensions in the last 12 months. 61% (43 out of 71) of the grievances pertained to irregular or delayed pension payments, while 34% (24) of individuals claimed having not received their monthly pension during the last 12 months. About 6% (4) respondents faced grievance related to paper work issues.

### From grievance to complaining and resolution

Table 2 presents the life cycle of pension related grievances - this helps us understand the working of the redress mechanisms, both at the level of financial service providers (FSPs) as well as the regulators. As described earlier, 622 respondents owned a pension product, while 71 had a grievance. Out of the 71, 59 (83%) complained to the FSP. The FSP was able to resolve 33 (56%)complaints. This implies that 26 complaints were not resolved. Of these only 8 (31%) were escalated to a higher authority, leading to a resolution of 5 (63%). Overall, this suggests that 33 grievances (46%) were not resolved - either because the respondent didn't complain at all, or because the problem was not resolved either at the FSP or regulatory level.

Table 2: Grievances, complaints and resolution
Pension Product N
Own the product 622
Complained to FSP 59
Resolved by FSP 33
Escalated to higher authority 8
Resolved upon escalation 5

We also explore the reasons why people do not complain when faced with a grievance with pension products. We focus on those respondents who didn't complain to the financial service provider/regulator. This doesn't include those who did not escalate their complaint after it was not resolved by the FSP. As seen in Table 2, 12 out of the 71 respondents who had a grievance chose not to complain. Of these 12 respondents, four felt that their problem would not get resolved sometimes because they didn't know if their problem was valid in the first place, while four were reluctant to access the process - either because they didn't have enough knowledge of the same, or because they felt the process was too costly and complex. These are very small sample sizes, and hence the results may not be generalisable.

### Impact of grievance on usage

In Table 3, we present the impact on usage for those who had faced any grievance while using the pension products.

Table 3: Impact on usage of pension schemes
Impact on usage of pensions N %
Changed the provider 36 51
Stopped using the product 11 15
No change 11 15
Reduced the use of product 5 7
Do not know/wish to answer 5 7
Increased the use 3 5

Regardless of the respondent's course of action, the experience of having faced a grievance is bound to have an impact on the usage. As a result of encountering grievance, a majority of the respondents chose to change their service provider. 51% of those who had a grievance (36 out of 71) changed their service provider. 7% (5 out of 71) users having faced grievance reduced the usage of the pension products, while 15% (11 out of 71)stopped using the product.

### Conclusion

GRMs in the pensions sector seem to be performing better than the banking and payments sectors. The incidence of grievances is lower, and the complaint rate is higher. However, banking and payments have a substantially larger number of users, and the products also get used more frequently than a pensions product. So it is not surprising that frictions in the banking space are higher. While the incidence of complaints may be lower in pensions, the impact of poor service may be higher on users, especially as the nature of grievances suggests that these occur later in life, when people may have limited means to solve the problem. This makes the numbers reported in the survey large enough to matter.

Vimal Balasubramaniam is a researcher at Queen Mary University, London. Aishwarya Gawali and Nancy Gupta are researchers at NIPFP. Renuka Sane is a researcher at Trustbridge. Srishti Sharma is a PhD student at Texas A&M University.

## Wednesday, January 11, 2023

### Coping with stress: Household borrowing for debt repayment

by Aishwarya Gawali and Renuka Sane.

There are concerns that household balance sheets are under stress due to rising levels of debt. An important contributor to the composition and levels of household debt is the ability to repay loans on time. Difficulty in repaying debt is also potentially a useful indicator of stress on the household balance sheets. There is already some evidence which indicates that households, and in particular rural households, are borrowing for debt repayment to cope with this stress. We present evidence on the same, from a large-scale panel survey.

### Data

Our analysis is based on data from the Consumer Pyramids Household Survey (CPHS), a pan-India panel household survey of about 174,000 households carried out by the Centre for Monitoring Indian Economy. In order to capture the latest available information from the CPHS, we use data from May to August (Wave 2) of 2022 and trace back to the same months of the preceding years. That is, we study data collected in the months of May to August (Wave 2) in each year between 2015 and 2022.

The data on sources and purposes of borrowing is sourced from the Aspirational India database within CPHS. This is our primary source of data for understanding credit access. Households are asked questions on their borrowing status across multiple sources and purposes. The responses are recorded as Yes/No, that is whether households have debt outstanding, and if so, whether households have borrowed from a specific source for a specific purpose. For example, if a household has borrowed for debt repayment, we would also get information on which source they borrowed from for debt repayment.

The data on income comes from the Household Income database which we use to create income deciles. The deciles are based on average monthly income of the five years prior to 2022.

We use household weights to get population level estimates of the share of households that have outstanding debt for the purpose of debt repayment.

### Q1: How many households borrow, and how many borrow for debt repayment?

Figure 1 shows the number of households with debt outstanding, and number of borrower households who have borrowed for debt repayment. The number of borrower households increased from 17 million in 2015 to 165 million in 2022. The number of households borrowing for debt repayment rose from just 0.75 million in 2015 to 23 million in 2022. The share of borrowing for debt repayment in overall borrowing rose from 4% in 2015 to 14% in 2022.

The pandemic year of 2020 is interesting because we find that there was a decline in the number of borrower households. However, there was an increase in the number of borrower households for debt repayment. Borrower households dropped from 154 million in 2019 to 141 million in 2020; however households borrowing for debt repayment increased from 12 million in 2019 to 14 million in 2020. As of August 2022, 165 million households in India had debt outstanding. Of these, 23 million had taken a loan to repay debt.

### Q2: What is the rural-urban variation in borrowing for debt repayment?

The CPHS data suggests that rural India has a larger share of borrower households than urban India. This is also borne out by the All India Debt and Investment Survey conducted by the NSSO, which shows that 35% of rural households are indebted compared to 22% in urban centres.

Figure 2 examines the rural-urban distribution of borrowing for the purpose of debt repayment. It shows two data points: the total number of borrower households in both rural and urban India, and the number of borrower households that have debt outstanding for reasons of debt repayment.

In 2015, there were about 16 million households in rural India and 9 million households in urban India that had debt outstanding. Of the 16 million in rural India, 2.26% had borrowed for debt repayment reasons. Of the 9 million in urban India, 1.35% had borrowed for debt repayment reasons. This number had increased to 12% of rural borrower households and 14% of urban borrower households in 2022. However in terms of absolute numbers, there are a lot more borrower households (overall and for debt repayment) in rural India than in urban India.

### Q3: What are the sources of borrowing for households with borrowing for debt repayment?

CPHS provides information on the source of borrowing for each purpose. Figure 3 presents the share of the main sources from which households have borrowed for debt repayment between 2015 and 2022.

There are two interesting patterns that emerge. First, in the earlier years, the biggest source of borrowing for debt repayment were the money-lenders. In 2015, 84% of households borrowing for debt repayment had borrowed from a moneylender. This has now flipped to self-help groups.

Interestingly, we also find that borrowing for debt repayment comes largely from Andhra Pradesh and Telangana. Perhaps the reliance on self-help groups is just a reflection of the largest share of borrower households for debt repayment being located in these regions which have a higher presence of self-help groups.

The next major source is relatives, friends and family followed by banks. It is important to note that these shares may not always add up to 100 because a household can take a loan for repayment of existing debt from more than one source.

### Q4: What is the relation between income and borrowing for debt repayment?

Figure 4 plots the number of households with borrowing for debt repayment, in each income decile for Wave 2 of 2015-2022. Borrowing is higher between the fourth and eighth income decile households.

Cumulatively, the fifth, sixth, seventh and eighth deciles accounted for almost 60% of the total number of households that borrowed for debt repayment in 2022. In the eighth decile, 3.93 million households borrowed for repayment in 2022. This indicates that 17% of the households that borrowed for debt repayment belonged to the eighth decile. 15% of the total household borrowing for debt repayment was from the sixth decile, with 3.47 million borrowers. The seventh and fifth deciles followed closely with shares of 14% (3.33 million) and 13% (3.10 million) respectively.

The number of households in these deciles have also grown over the years and 2022 has the highest number of households in each decile. The lines for 2015 and 2016 are almost flat which indicates the low level of borrowing for debt repayment. The rise becomes prominent from 2017 as the gaps between the lines start to widen. This is most visible in the case of 2022, where all deciles, but especially the sixth and eighth decile saw a huge rise in borrowing for debt repayment.

### Conclusion

In this article we have established some basic facts about household borrowing for debt repayment in India. First, there has been an increase in the number of borrower households, as well as the number of households who borrow for reasons of debt repayment. Second, the increase is greater in rural areas as against urban areas. Third, self help groups seem to be the most prominent source of borrowings for debt repayment. Fourth, households between the fourth and the eighth decile have the highest number of households with borrowing for debt repayment. These facts are building blocks of a larger research agenda on understanding debt and distress.

Renuka Sane is a researcher at Trustbridge and Aishwarya Gawali is a researcher at NIPFP.

## Wednesday, December 28, 2022

### Examining grievances and redress for banking products

by Vimal Balasubramaniam, Aishwarya Gawali, Renuka Sane and Srishti Sharma.

Banks are witnessing persistent consumer complaints. These range from high service charges, lack of transparency in pricing and mis-selling. The regulatory system is currently not designed to capture the total number as well as nature of grievances. Without this information, it becomes difficult to design a policy solution.

We examine the grievance and redress experience for a heterogeneous set of consumers, with a large scale survey in five major states of India. Through this survey, we map the journey of a consumer's experience-- from usage and grievance to resolution. We also study the impact that a grievance has on subsequent usage of the product. In this article, we present information on the consumer's experience for a comprehensive set of banking and payment products that fall under the regulatory ambit of the Reserve Bank of India. These include:

1. Banking deposits.
2. Bank credit.
3. ATM/Debit Cards.
4. Net banking/ Phone Banking(including NEFT/IMPS/RTGS).
5. NBFC.
6. UPI Wallets.
7. Microfinance institutions.
8. Co-operative credit societies.

The evidence from our study can form the basis of a more responsive system of grievance redress for retail consumers.

### The survey design

Our survey was conducted in Maharashtra, Bihar, Haryana, Madhya Pradesh and Andhra Pradesh. We used a multi-stage stratified sampling method to draw the sample of households. The Primary Sampling Units (PSUs) were the villages for rural areas and census enumeration blocks (CEBs) for urban areas. The Ultimate Sampling Units (USUs) were the households from these PSUs.

The 2011 Census served as the sampling frame for the identification of the districts within each state. All the districts in a state were divided into terciles on the basis of distribution of households using banking deposits, curated from the RBI data across four quarters of 2020-21. To ensure proportionate distribution in each tercile, two districts were picked from each tercile using systematic random sampling. This exercise was repeated for each of the five states. In states such as Maharashtra, Bihar & Haryana where one district held a substantially high proportion of deposits, it was treated as two districts and over sampled to account for the large proportion of deposits. Accordingly, we got a sample of six districts each from Madhya Pradesh and Andhra Pradesh, and five districts each from Maharashtra, Bihar and Haryana which gave us a total of 27 districts.

Within the district, the allocation of the sample between villages and CEBs was proportional to the rural-urban distribution of the population. Villages in a district were stratified on the basis of distance from district headquarters and CEBs were stratified on the basis of share of CEB in the district's urban population. Three strata were created on the basis of the above mentioned criteria, for both villages and CEBs. The number of households from each strata was selected in proportion with the population share of each strata. So if one village strata has 40% of the rural population, then 40% of the rural sample of the district came from that strata.

We collected information on the demographics, physical and financial assets, and liabilities of the household. The core module of the questionnaire focused on experience of consumers with grievances & redress regarding financial products. The total sample size was 21,355 respondents.

### Measuring grievances in the banking system

According to the current grievance redress system for banks, NBFCs and Prepaid Payment Instruments, consumers must first lodge a complaint with their service provider. If the service provider is unable to provide resolution in a satisfactory manner in 30 days, the consumer may escalate the complaint to the Ombudsman. The Ombudsmen for banks, NBFCs and digital payments have been harmonised under the Integrated Ombudsman Scheme in 2021. Consumers can lodge their complaints with the Ombudsman using the Complaint Management System (CMS) portal or by using a complaint form. We asked the following questions to understand the consumer's experience at each step of the grievance redress process:

• First we asked if the respondent is using/has ever used the mentioned product. This ensures that we also capture past users of a financial product.
• We asked if they have faced an issue with the mentioned financial product in the last 12 months. This helps us capture the grievances faced by consumers.
• We then asked the respondent what their latest/most recent issue was for the mentioned financial product.
• We also asked them their first course of action after encountering the grievance. Through this question, we understand how many people complain after facing a grievance. We consider that a respondent has complained if they went to the service provider or regulator with their issue. We do not include complaints to the police or consumer courts as this is not in the ambit of the regulatory grievance redress system.
• Additionally, we asked if their issue was resolved after their first complaint.
• For those who did not receive resolution at this stage, we asked if they escalated the complaint to a higher authority.
• For those who escalated their complaint, we asked if they finally received resolution.
• To all those who faced a grievance, we asked what the impact of the grievance was, on their usage of the product.

### Overview of grievances in the banking system

In Table 1, we describe the extent of usage and grievance for the various banking and payment products. Columns (1) and (2) provide the number and percentage of the sample who have used a particular financial product. Columns (3) and (4) describe the number and proportion of users that reported that they faced an issue/grievance related to the financial product in the last 12 months.

Table 1: Usage and incidence of grievance
Product Usage Incidence of grievance
(1) (2) (3) (4)
N % N %
Banking deposits 17407 81.51 2112 12.13
ATM/Debit Card 8625 40.39 1279 14.83
Netbanking (IMPS/NEFT/RTGS) 3161 14.80 503 15.91
UPI Wallets 2825 13.23 531 18.80
Bank credit 1640 7.68 242 14.76
MFI 961 4.50 104 10.82
NBFC 448 2.10 82 18.30
Cooperative credit society 386 1.81 72 18.65

Banking deposits were the most used product, followed by ATM/Debit cards. Netbanking was the third most used product, followed closely by UPI wallets. All the others were used by less than 10% of the sample each. Co-operative credit societies had the smallest share of users.

The incidence of grievance ranged between 10% to 19%. Even though bank deposits were the most used product, they had the second lowest incidence of grievances. The highest grievance rate was for UPI wallets, at about 19%. Co-operative credit societies had the second highest incidence of grievances, even though the usage of these products was the lowest in our sample. The same is true for NBFCs as well - only 2% of the sample used NBFC products, but 18% of these had faced a grievance.

### Nature of grievance

In Table 2, we describe the nature of grievances faced by consumers. We select the top 3 grievances for each product and present the number and proportion of consumers who faced the given issue.

For banking deposits, 28% of those with a grievance had an issue related to transaction failure. 26% of the issues were related to charges being deducted without information. More worryingly, 12% of the issues were related to difficulties with opening a bank account. This has implications for financial inclusion as such issues may dissuade people from participating in the formal financial system.

 Product N % Banking deposits Failures or delays in transactions 585 28 Deductions or charges without information 555 26 Difficulty in opening bank account 244 12 Bank credit Loan/interest rates 130 54 Fraud, hidden charges etc 102 42 Other, specify 10 4 ATM/Debit card Lack of cash in ATM 490 38 ATM closed/non functional 259 20 Server down 156 12 Netbanking Server down 204 44 Delays in services 92 20 Money deducted but transaction failed 85 18 UPI Wallets Server down 188 35 Delays/Failure of transactions 115 22 Transaction failed, but money debited 99 19 NBFC Complex terms and conditions/no adequate notice 36 44 Non transparency in contract/ loan 23 28 No communication about loan sanctioned 7 9 MFI Charged higher interest rate than informed 56 54 Threat to increase interest rates 27 26 Painful recovery process 12 12 Co-operative credit society Charged higher interest rate than informed 42 58 Threat to increase interest rates 23 32 Painful recovery process 6 8

For bank credit, the dominant issues were interest rate related, however, 42% of consumers reported having faced fraud which is a far more serious nature of grievance.

For ATM cards, 38% of the issues are related to lack of cash in ATMs, 20% were related to dysfunctional ATMs and 12% of the issues came up because the server was down. 44% of netbanking issues were also because the server was down. While transaction, server and service infrastructure related issues dominate in case of payment products such as UPI and net banking, interest rate and loan contract related issues are significant for credit products.

### From grievance to complaints and resolution

In Table 3, we report the number of people who complained to either the service provider or regulator when faced with the grievance. Column 2 presents the total number of grievances for the given product. In Column 3, we report the number of users who complained to either their service provider or regulator. In Column 4, we report the number of consumers whose complaint was resolved at the first stage itself. In column 5, we report how many of those who escalated their complaint reported that their issue was resolved.

 (1) (2) (3) (4) (5) Product Had a grievance Complained to FSP Resolved by FSP Escalated to higher authority Resolved upon escalation Banking deposits 2112 1064 661 88 47 Bank credit 242 201 104 61 6 ATM/Debit Card 1279 521 410 39 22 Netbanking(IMPS/ NEFT/RTGS) 503 252 180 24 13 NBFC 82 72 43 8 4 UPI Wallets 531 187 141 9 6 MFI 104 69 34 8 4 Cooperative credit society 72 64 37 7 5

Out of the 2112 consumers who faced a grievance regarding banking deposits, 1064 (50%) complained, which is how we define the complaint rate. Of these, 661 (62%) reported that their problem was resolved after their first complaint, which is how we define the resolution rate. This suggests that 403 users' complaints were not resolved in the first instance. Of these, only 88 (21%) escalated the complaint. Of these 88, 47 (53%) reported that their problem was resolved after escalation. The other half of the complaints that were escalated remain unresolved.

Banking deposits have the highest number of complaints. However, co-operative credit societies had the highest complaint rate at 88% followed by NBFCs at 87%. ATM card complaints have the highest rate of resolution at the first stage -- 410 out of 521 complaints (78%) got resolved at the first stage. This is followed by UPI wallets at 75% and netbanking at 71%. High resolution rates suggest that grievance redress at the first point of contact, which is usually the Financial Service Providers (FSPs) is performing efficiently. MFIs have the lowest resolution rate (49%) at the first stage, and points to the deficiencies in the redress system.

### Reasons for not complaining

In the previous section we examined what happens to the complaints that enter the official grievance redress system. However, it is evident that not all grievances turn into complaints. What about the users who do not lodge a complaint?Table 4 shows the reasons why people do not complain when faced with a grievance. The rows show the number of users who did not complain for the reason given in the column.

For banking deposits, the main reason for not complaining was that users did not know the process of grievance redress. 38% of those who did not complain, did so because they did not know the process. For bank credit, the costly and complex nature of the process was the main reason for not complaining with 34% users not complaining due to this reason. 36% ATM/debit card users who did not complain did so due to the complicated and expensive nature of the grievance redress process. 26% did not complain because they did not know the redress process. Another 15% did so because they were not sure about whether their problems would be resolved. For netbanking and phone banking, 33% users did not complain because the redress process is too costly and complex, 15% did not complain because they were not sure about their issue being resolved and 17% because they didn't know the process.

 Product Did not complain to FSP Costly and complex process Did not know validity of complaint Do not know/wish to answer Fear of retribution Resolution unlikely Unknown process Was advised not to by friends family Banking deposits 1048 177 88 96 38 188 408 19 Bank credit 41 14 6 - 2 6 5 2 ATM/Debit card 758 278 71 31 17 120 198 17 Netbanking 251 85 25 18 7 43 39 12 UPI wallets 344 114 41 9 5 60 94 15 NBFC 10 1 - - 2 2 2 - MFI 35 - 1 3 2 8 19 - Co-operative credit society 8 - - 1 1 2 -

In the case of UPI wallets, the expensive nature of the process, lack of information about redress procedures and the prospect of resolution being unlikely, were the main reasons for not complaining. For NBFCs, 20% of users who did not complain, did so because of the fear of retribution, another 20% did not complain because they were not sure about whether their problems would be resolved. Finally, 20% did not complain because they did not know the process. For MFIs, 54% of users did not complain because they were unaware of the process and for co-operative credit societies, this number was 25%.

### Impact of grievance on usage

The experience of having faced a grievance is bound to have some impact on the consumer's usage of the product. In Table 3, we present the number and proportion of people who either changed their provider, reduced usage of the product or stopped using the product after facing a grievance. These actions indicate that the grievance had an adverse impact on the user. Grievances related to deposit or payment products lead about 30% of consumers to take some action. The response by consumers is higher for credit related products.

Table 4: Changed provider/reduced/stopped usage
Product Had a grievance Took action
(1) (2) (3) (4)
N N %
Co-operative credit society 72 63 88
NBFC 82 65 80
MFI 104 65 62
Bank credit 242 129 52
Netbanking 503 166 34
UPI Wallets 531 173 31
Banking deposits 2112 666 31
ATM/Debit card 1279 388 29

In the case of co-operative societies, 88% of those who faced a grievance either changed their provider or reduced or stopped using the product as a result of it. This indicates the co- operative society members who faced a grievance may not have had a satisfactory experience with the grievance redress process. This number stands at 80% for NBFCs and at 62% for MFIs. The impact of grievances for non-bank lending institutions is far more adverse than for any of the other products.

### Conclusion

Improving outcomes for consumers is one of the core goals of finance. It is important to understand how the system deals with grievances of consumers, and where there is scope for improvement. Our results present a heterogenous picture. We find that the usage of deposit and payment products is higher than credit products, while the grievances are higher for credit products. Deposits are the most used product but have an incidence of grievance of 12%. Co-operative societies are used by less than 2% of the sample, yet almost 19% of its users have faced a grievance. NBFCs are used by just about 2% of the sample, and have an incidence of grievance of 18%. Deposit and payment related services are able to resolve grievances faster - more than 70% of the complaints were resolved in the first instance. This is not true of credit related products - for example, only 49% of the complaints were resolved for MFIs. Further research could explore the possible reasons for this heterogeneity.

Renuka Sane and Aishwarya Gawali are researchers at NIPFP. Vimal Balasubramaniam is a researcher at Queen Mary University, London. Srishti Sharma is a PhD student at Texas A&M University.

## Sunday, December 25, 2022

### Author: Pavithra Manivannan

Pavithra Manivannan is a Senior Research Associate at XKDR Forum and Chennai Mathematical Institute.

## Friday, December 23, 2022

### Delays in government contracting: A tale of two metros

A state entity undertaking a procurement exercise must meet prescribed timelines throughout its procurement pipeline. Delays in one or more milestones adversely affect all parties involved: the procuring entity (increase in expenditure beyond the budget and disputes), the contracted vendors (uncertainty and delays in payment) and the public (delays in utilising public goods and services). At the outset, we recognise that the indicator of a successful procurement exercise are multi-fold: achieving required quality, adhering to timeline and limiting spending gap. Our approach employs the lack of delays as the indicator of a successful procurement exercise.

In a recent article, we examined the extent to which DMRC's (the Delhi Metro Rail Corporation) competence in timely project execution was borne out by data. We found that (a) DMRC is able to meet the Government of India's and its own stipulations in two stages of its procurement process, that is, contract award and vendor payments; and (b) In spite of this exemplary performance, DMRC has faced delays in overall project implementation that have gradually increased over time. This article seeks to understand the underlying factors that potentially contributed to DMRC's prompt performance in its procurement process.

It is not possible to understand DMRC's success in isolation. Instead, we analyse it relative to its predecessor, the Calcutta metro-rail system (Calcutta metro). The Calcutta metro was India's first metro-rail system to be implemented. It was plagued by delays and cost overruns. Such a comparative analysis of a successful public project to one that fared worse in execution is revealing. First, it shows the learning curve of the state in building capacity to execute public projects. Second, it helps to understand what works and what does not, when a state entity conducts a procurement exercise. The analysis can serve to provide valuable feedback in procurement reform policies.

### Delays in execution of metro-rail systems

Formerly, the Ministry of Railways was responsible for the construction of mass rail services, including metro-rail systems in metropolitan cities. The Ministry undertook the construction of the Calcutta metro in 1971. In 1986, the Government of India (Allocation of Business) Rules, 1961 was amended to shift the responsibility of the urban transport system to the Ministry of Urban Development (now the Ministry of Housing and Urban Affairs). In contrast to the Calcutta metro, the construction of the Delhi metro-rail system (Delhi metro) was undertaken by the Ministry of Urban Development as the nodal union ministry. The role of Railways was limited to providing technical assistance.

We study the annual reports of the Delhi metro and the Parliamentary Public Accounts Committee Reports (1981; 1989; 1992) on the Calcutta metro, to estimate overall delays in both these projects. We source this data from the website of DMRC and the Parliament of India, respectively. Our data consists of six time periods during which there was significant procurement of works. The data includes the date of completion and submission of the detailed project report (DPR), the date on which the project received Union Cabinet approval, the date of commencement of works, the scheduled date of completion of the project, and the actual date of completion of the project, in part and full. These are presented in Table 1 below, as a timeline of events for the first line of the Calcutta metro and the first phase of the Delhi metro.

Table 1: Timeline of events
Event Calcutta Metro Line-1 Delhi Metro Phase-1
Completion of Detailed Project Report (DPR) 1971 1995
Project sanction 1972 1996
Project commencement 1978 1997
Scheduled completion 1978 2005
Partial commission (one section) 1984 2002
Project completion (fully operational) 1995 2006

We find that the gap between estimated and actual date of completion is a little more than a year for the Delhi metro. This gap was close to two decades for the Calcutta metro. Further, the lag between the date of sanction of the project to the date of commencement of works for the project is wider for the Calcutta metro (4 years) than for the Delhi metro (about a year).

This suggests that, from 1971 to 1995, there appears to be much improvement in the way procurement was undertaken for Indian metro-rail systems. We posit that the Delhi metro's success was shaped by the challenges faced and the experiences gained in implementing the Calcutta metro. Our analysis attributes learnings from the Calcutta metro to the following structure and list of processes adopted by DMRC: its institutional design, its financing and revenue models, global transfer of technical know-how, and expertise of its early leadership. In the subsequent sections we analyse how each of these features enabled the Delhi metro to avoid inordinate delays.

### The institutional design of the procuring entity

What motivated the institutional design of DMRC? To answer this we look at various Parliamentary Committee reports, CAG reports and literature on the subject. Our review suggests that there were three main institutional constraints faced by the Ministry of Railways in implementing the Calcutta metro.

1. The lack of coordination with the West Bengal State Government and the local agencies in Kolkata. There were delays in land acquisition, problems in utility diversions such as transport, water and sewage, and detection of uncharted utilities after commencement of works. These instances had a direct impact on the contracting process, such as frequent interruptions of works, revisions to scope of work, and change in construction methodology (Public Accounts Committee, 1981).
2. Frequent changes and vacancies within the Ministry of several important personnel such as the General Manager and Chief Engineer. This was due to the administrative process of the Ministry. The Railways had to follow the conditions laid down by the Appointment Committee of the Cabinet with respect to retirement, superannuation and promotion (Public Accounts Committee, 1981). This resulted in loss of experience and expertise within the procuring entity.
3. Inadequacy of financial powers delegated to the General Manager. From the year 1974 to 1982 the General Manager had the power to sanction tenders up to Rs. 1 crore only. This was increased to Rs. 2 crores in 1983 and Rs. 5 crores in 1985. This limited power of the General Manager meant, approvals for sanctions of higher value tenders had to be received from the Railway Board. This procedure was time consuming and caused delays in finalisation of contracts by up to 3 years (Public Accounts Committee, 1989).

We speculate that the above constraints prompted the authorities undertaking the Delhi metro project to adopt a different approach. The Calcutta experience provided two guides for the organisational structure of DMRC. One, to build institutional capabilities for executing a metro-rail system outside the Ministry of Railways. Second, to have a separate corporate entity with independent decision making authority. Thus, DMRC was set up with two distinguishing features which worked in its favour: It was formed as a limited liability company under the Companies Act, 1956 and the ownership of the entity vested equally in the Union and the State Government. The board of directors of DMRC constituted representatives from the Union Ministry of Urban Development, Department of Transport of GNCTD and the Delhi Development Authority. Such an institutional arrangement, by aligning incentives for all the stakeholders, enabled better coordination with the local government and ensured that the management had the backing of both the State and the Union Government. Further, functional directors appointed for distinct functions such as, project and planning, works, electrical, finance, business development and the like, had sufficient powers delegated to them under the Schedule of Powers (CAG, 2008). This facilitated quick decisions in expenditure approvals, qualification of bidders, finalisation and acceptance of contracts. Finally, the long tenure of key personnel such as the Managing Director, enabled the organisation to retain domain experience and expertise.

### Financing and revenue models

For prompt execution and sustenance of any infrastructure project, timely flow of funds is essential. Metro-rail systems are capital-intensive projects. The Calcutta metro was fully funded by the Ministry of Railways. One of the main reported reasons for delay in the project was lack of funds and improper utilisation of allocated funds. Up until 1980, the Railways had not fully utilised the funds allocated for the project. Further, for subsequent years, sufficient funds were not made available for the construction. This resulted in shortage of raw materials, such as steel and signaling equipment, and delayed payments to vendors (Standing Committee on Railways, 1993; Public Accounts Committee, 1981). For the DMRC project, the Calcutta experience prompted the authorities to explore other avenues for funding such as, equity, external agency loans, subordinate loans from centre and state, property development revenue and central government grants. Most significant was the official development assistance (ODA) loan from the Japan International Cooperation Agency (JICA). Nearly 54-55 per cent of the first three phases of the DMRC projects was funded by JICA as a low-interest and long-term concessional loan. The funding pattern for each phase of the project sourced from the DMRC website is as set out in Table 2. Smooth flow of funds into DMRC enabled timely payment to vendors and ensured that the project was not delayed due to uncertainty in financing.

Table 2: Funding pattern for DMRC projects
Phase I Phase II Phase III
JICA loan 60% 54.47% 48.57%
Equity from GoI 14% 16.39% 10.04%
Equity from GNCTD 14% 16.39% 10.04%
Loans from Union/States 5% 6.56% 13.39%
Grants from States - 0.59% 10.62%
Property Development 7% 5.59% 7.34%

Another lesson came from the fact that the Calcutta Metro was not financially viable (Singh, 2002). The traffic earnings were inadequate to cover the operating expenses of the metro-rail system. This not only burdened the exchequer in the form of subsidies but also affected the public as the fare per trip that was charged had to be increased to sustain operations (Public Accounts Committee, 1989). Decrease in cash flow meant stalling of procurement of raw materials, and delays in payments to vendors.

The financial crunch faced by Calcutta Metro encouraged DMRC to generate revenue through non-conventional sources. DMRC adopted the examples of well-performing international metro-rail systems and sought to increase its non-farebox revenue. Table 3 below shows the revenue model of DMRC for 10 years (FY2011-FY2020) obtained from its annual reports. Revenue from traffic operations is categorised as fare-box revenue and revenue from real estate, consultancy and external projects are categorised as non fare-box revenue.

Table 3: DMRC Revenue Model (as percentage of total revenue)
Description Fare revenue Non fare revenue
FY20 65.49 34.51
FY19 62.92 37.08
FY18 55.22 44.78
FY17 45.69 54.31
FY16 53.35 46.65
FY15 60.34 39.66
FY14 55.74 44.26
FY13 62.93 37.07
FY12 65.74 34.26
FY11 65.05 34.95

On an average 58.49% of DMRC's revenue is from traffic operations (fare-box revenue) and 41.51% of the revenue is through other sources (non-fare box revenue). This is in line with international practice. For instance, the non-farebox revenue of some of the better performing metros in the world (in terms of ridership and network length), such as London, Singapore and Hong Kong, ranges from 25-60% of its total revenue. DMRC's capacity to source funds and remain financially viable has helped it to make timely payments to its contractors, repay its debts, and expand its network line.

### Human capacity and technical know-how

Building human competence within the government is paramount to do procurement well. This includes both functional as well as technical competence. In India, the technical know-how to build metro-rail systems was lacking. The Calcutta metro was the first ever underground railway project undertaken in India. Despite this, global tenders were not invited for construction of the work. Neither the construction firms in the country nor the Railway Administration possessed the experience to construct underground structures for a rapid transit system. The lack of expertise led to frequent abandonment of works and changes in scope of work, resulting in huge financial implications in addition to time overruns (Public Accounts Committee 1989 and 1992). Thus, when the idea of a metro-rail system in Delhi was born, the need to rope in personnel with prior expertise and experience, such as, B.I. Singal and E. Sreedharan, was recognised.

Mr. B.I. Singal was the former Director General of the Institute of Urban Transport and the then Managing Director of RITES (Rail India Technical and Economic Service). Mr. Singal came in with 11 years of experience in the planning and building of some of the finest metro-rail systems in the world, such as the Hong Kong MTR (known for completing the project within time and budget) and Taipei metro-rail network. RITES prepared the feasibility study on building a metro-rail system for Delhi. Mr. Singal made sure that his RITES team had a few professionals who had previous experience of working with the Calcutta metro. Mr. E. Sreedharan, the first Managing Director of DMRC, had served as the Chairman and Managing Director of Konkan Railways. He brought in his domain experience of working with the Railways as well as the management experience of heading an autonomous entity. Studies document some effective practices adopted by Singal and Sreedharan which we speculate had an impact on the organisation's procurement practices. They insisted on independence in decision making, speed, and global exchange of knowledge and expertise (Ashokan, 2015; CPI, 2017). This resulted in creation of DMRC as a separate legal entity and in transfer of Japanese technology and know-how in building metro-rail systems.

After the failed attempt at indigenisation by the Calcutta metro, the authorities felt the need to tap in to global expertise for the Delhi metro project. In addition to funding from JICA, Japanese Consultants were also brought on board. This ensured transfer of foreign technological knowledge, skills and expertise to DMRC. DMRC engineers developed technical skills such as tunneling technologies, and functional skills such as management ethos, and value for time from their Japanese counterparts (Onishi, 2016). This enabled DMRC to build in-house capacity, which now helps other metro-rail networks in the country.

### Discussion

Our work shows how the Indian state attempts to achieve better outcomes by identifying lessons from its past shortcomings. The challenges faced by the Calcutta metro shaped the Delhi metro's institutional design, financial structure, and human resource competence. Our article highlights the importance of these three factors in enabling desirable procurement outcomes.

A key insight from our analysis is that these factors do not work in isolation. Autonomy in decision making, efficient and experienced personnel, adequate financing, and right institutional choice are all inter-operable and go hand in hand. If a procuring entity seeks to realise better outcomes, procurement reforms must not merely pick the lowest hanging fruit of these factors. Instead, a sector-specific approach of studying the past experiences must be employed to act as feedback into future projects. Our research provides a framework to assess such past successes and failures, and demonstrates the potential of deploying such research.

### References

Public Accounts Committee, Fifty-fifth Report, 1981, Hundred and Forty-second Report, 1989 and Ninth Report, 1991.

Standing Committee on Railways, Second Report, 1993 and Thirty-fourth Report, 2007.

Comptroller and Auditor General of India, Report No. Performance Audit 17, 2008.

Pavithra Manivannan, Lessons from the Delhi Metro, Business Standard, July 2021.

Anirudh Burman and Pavithra Manivannan, Timeliness in government contracting: Evidence from the country's largest metro-rail network, The Leap Blog, August 2022.

Yumiko Onishi, Breaking Ground: A Narrative on the making of Delhi Metro, JICA, 2016.

Centre for Public Impact, The Construction of the Delhi Metro, November 2017.

Saurabh Singhal, Non Farebox Revenue for Metro - A Global Perspective, Business World, May 2022.

The International Association of Public Transport (UITP), World Metro Figures, 2018.

M.S. Ashokan, Karmayogi - A biography of E. Sreedharan, Penguin, 2015.

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.

## Monday, November 28, 2022

### The litigant perspective upon courts

How do we identify a well performing judiciary from one that does not perform well? The literature on this question has focused on two types of metrics: inputs such as the judge to population ratio, judicial budgets and physical infrastructure and outputs such as the number of resolved cases, time taken per case and the costs involved. An emergent literature focuses on the litigant's experience of the judiciary. This approach involves criteria that the litigant uses to evaluate their experience of the judiciary, which have been found to be different from those used by judges, legal practitioners and planners (eg., Tyler, 1984; Rottman and Tyler, 2014; Hagan 2018).

In India, there is a growing awareness for the judiciary to be more citizen friendly (example; example; example), which calls for a better understanding of what a litigant's expectations are when engaging with the judiciary. In a new working paper, we propose a measurement framework that focuses on the litigant's perspective. In order to construct the framework, we draw upon the literature to hypothesise what a litigant takes into consideration when she decides to take a dispute for adjudication at a court. These considerations are then translated into the metrics to be used, when designing an evaluative framework to compare courts with similar functions. When this framework is applied to data from the legal system, it becomes an information system which can generate quantitative expectations of the time and costs involved in the process of litigation, which can potentially guide the litigant on whether to litigate.

In designing such a measurement framework, we recognise that there cannot be a single set of metrics that can be applied equally to all courts. This is because different courts perform functions that vary substantially in complexity, type and processes. For instance, the evidentiary burden required to be followed in a criminal matter is different than that of a civil matter, and the prosecution is led by the state. Additionally, the intended relief to a litigant in different types of matters also varies. For instance, in a civil matter, the relief is largely limited to compensation, specific performance and/or damages from the defendant. On the other hand, in constitutional matters, the relief sought may involve directions to the state or lower courts. While there may be some common metrics that could be useful to evaluate different types of courts, a single set of metrics may make the evaluation framework over expansive or deficient for some types of courts. Therefore, in this paper, we limit the scope of our discussion and the resulting framework to courts that adjudicate contractual disputes.

#### Features of the proposed framework

Given the focus on contractual dispute resolution, we identify a list of five metrics from the literature which can be usefully applied by a litigant to evaluate the performance of a court. The metrics are independence, efficiency, effectiveness, predictability and access. Based on the multiple interpretations of each metric available in the literature, we present arguments that justify why we narrow down on one interpretation over another from a litigant's perspective. We then identify proxies that can be used in the Indian context to measure the performance of the chosen courts on the selected metrics. These make up the proposed framework to measure the performance of courts that adjudicate contractual disputes.

The metrics, and the proxies that can be meaningfully evaluated to assess the metric, and the description of each proxy are summarised in the Table below. Finally, in the paper, we also lay out the source of the data and the process in which the information on each of these metrics can be collected.

Table: Metrics for evaluating court performance on contractual disputes

Sr. No. Metric Proxy Description
1. Independence Procedural fairness Adherence to procedure
Distributive fairness Fairness and impartiality in judgements
2. Efficiency Timeliness Duration of disposed and pending cases
3. Effectiveness EnforceabilityRatio of sum recovered to the total sum awarded in court orders
4. Predictability Certainty of case trajectoryClarity on stages of the case
Hearing date certaintyCertainty on number of hearings per case
Ratio of substantive to non-substantive hearings
5. Access Monetary costsCosts of approaching the court to the litigant
ConvenienceEase and user-friendliness for litigants

There are two caveats to the measurement framework that we propose. First, we assume that the litigant assigns equal weights to each of these metrics in making her decision on whether to take a contractual dispute in court. This means, that the litigant values (say) independence as much as predictability. This is a simplification and may not necessarily hold in reality, and for each litigant. Second, we do not identify an optimal or ideal level of performance of the court on these metrics. For example, we do not attempt to identify an ideal duration for the disposal of a case or the optimal number of hearings or the optimal 'level' of independence. The aim of the proposed framework is simply to provide a transparent base of metrics about court performance that can be put together using publicly accessed data sources, that we believe matters to the litigant.

The public domain nature of the data used in the proposed framework, supports regular updates of the metrics. This, in turn, will facilitate a comparison of the performance of court adjudicating contractual disputes over time. If these measures can be calculated in a consistent manner across different platforms, these can provide the litigant with a relative performance evaluation that can allow her to decide when, if and how to avail of the justice delivery system with greater clarity and certainty.

#### Conclusion

While judicial under performance is an over used expression in both the academic literature and broader policy discourse on Indian courts, the absence of an evaluative framework exacerbates the ambiguity associated with this expression. Our literature review in this paper shows that what is measured in the context of courts largely depends on who is undertaking the measurement. By considering specific metrics that a litigant may attach priority to in her experience with the judiciary, this paper provides a foundation for rolling out regular evaluation exercises of courts adjudicating commercial disputes, and ultimately make judicial performance a more tangible and usable concept in India.

#### References

Hagan MD (2018). “A Human-Centered Design Approach to Access to Justice: Generating New Prototypes and Hypotheses for Intervention to Make Courts User-Friendly.” Indiana Journal of Law and Social Equality, 6(2), 199–239.

Rottman DB, Tyler TR (2014). “Thinking about judges and judicial performance: Perspective of the Public and Court users.” Onati Socio-legal Series.

Tyler, Tom R. "The Role of Perceived Injustice in Defendant's Evaulations of their Courtroom Experience." Law & Society Review, vol. 18, no. 1, 1984, p. 51-74.

Pavithra Manivannan is a senior research associate at XKDR Forum, Mumbai. Susan Thomas is Senior research fellow at XKDR Forum, Mumbai and Research Professor of Business at Jindal Global Business School. Bhargavi Zaveri-Shah is a doctoral candidate at the National University of Singapore.

## Saturday, November 05, 2022

### Learning by doing and public procurement in India

by Aneesha Chitgupi, Abhishek Gorsi, and Susan Thomas.

### Introduction

State objectives can feature a mix of make' (i.e. build an organisation, recruit people) vs. buy' (i.e. contract to a private person for the same task). While private persons are, in general, more efficient than a comparable government organisation, the `buy' pathway is not a panacea when there are bottlenecks in state capacity to contract. Understanding how to achieve capacity in government contracting is one of the critical components of the overall question of enhancing state capacity in India.

How has this tradeoff evolved over time? Generally, when contracting capabilities go up, the fraction of work that is contracted-out goes up. Chitgupi and Thomas (2022) find that the government procurement spending has been at a steady 17 percent of the total expenditure across multiple years, which suggests a lack of movement on contracting capabilities.

What about cross-sectional variation in contracting capability between spending units? Sharma and Thomas (2021) finds that procurement expenditure varies significantly across ministries in 2016-2017, with a few ministries accounting for the larger share of procurement. This would, of course, largely reflect differences in the nature of their work.

In this article, we obtain insights into contracting capabilities by examining procurement under-spend across ministries and across time. We recognise that state capacity in procurement includes multiple dimensions: achieving a targeted level of quality at the minimum cost and on time. Our question is a more basic one about whether the spending happens at all. It requires a certain level of contracting capability in order to actually achieve the budgeted procurement expenditure. We then consider the role of learning by doing in procurement.

An insight in the field of state capacity is that it is extremely sticky; changes to state capacity arise only slowly (Kelkar and Shah, 2022). State capacity emerges from the organisation design (mandate, governance, organogram, processes, the feedback loops of accountability mechanisms). Repetition of the same task, over and over, under the influence of accountability mechanisms, induces the slow process of building capability. By this reasoning, we hypothesise that state entities which procure on a sustained basis will build the organisation design and capacity to doing procurement efficiently, and vice versa. The lack of sustained procurement activity will then predict procurement underspend.

### Approach

In previous work, we developed the methodology to estimate the fraction of procurement in the accounts reported in the Detailed Demand for Grants (DDG) document published by a ministry. The DDG for a given year reports two sets of numbers by item head: the amount budgeted for the year, the amount actually spent two years previously. So, the DDG in 2016 for (say) the Ministry of Health and Family Welfare will report the budgeted amounts for 2016, as well as the amount spent in 2014. These are reported by item head, and allows us to calculate the actual amount spent but with a lag of two years.

For our current analysis, we calculate what was budgeted and what was actually spent on procurement, across ministries and across years. We use these to estimate the procurement spending gap for the ministry as:

100 x (actual procurement spend - budgeted procurement) / budgeted procurement

A negative spending gap has the ministry spending less than planned. If procurement capacity is the dominant factor driving the ability to procure, then we would say that a ministry with good procurement capacity will have a procurement spending gap close to zero. Further, if the learning-by-doing hypothesis holds, then those ministries that budget consistently for procurement will have built procurement capacity, and be able to show procurement spending gap close to zero.

Other than procurement, a ministry spends on operational expenses like salaries and pensions, and on grants-in-aid and schemes. These do not depend upon specialised capacity as with the contracting capacity required in the case of procurement spending. As a benchmark, we also calculate the total spending gap as the difference between the actual and budget total spending of the ministry, as a fraction of the total budgeted spending. We use the total spending gap of the ministry to benchmark their overall spending capacity.

We calculate the total spending gap of the ministry as:

100 x (actual total spent - total budgeted) / total budgeted

For our analysis, we do the following steps:

1. Estimate the amount of procurement expenditure for each ministry for each year for which we are able to obtain the DDG. We use this to identify which ministries have consistently higher fraction of their expenditure in procurement.
2. Estimate the procurement under-spend for each ministry across different years.
3. Examine whether there is a link between the ministries with a higher fraction of procurement expenditure and a relatively lower procurement under-spend.

### Data

We collect the DDG published by the ministries of the Union Government to estimate the procurement and total spending gaps. The estimation procedure described in Sharma and Thomas (2021) requires the object head wise expenditure to be disclosed in the DDG. This level of information has been consistently disclosed across all ministries (other than for the Railways ministry which has a more complicated accounting structure) from 2014-2015 onward. The latest available DDGs are from 2020-2021.

There remains challenges in accessing the DDGs reliably for all the Union government ministries. In our exercise, we are able to locate the DDGs consistently from 2014-15 to 2020-21 for 10 ministries: Civil aviation, Coal, Environment, Finance, Health and Family Welfare, Home Affairs, Housing and Urban Affairs, Rural Development, Road Transport and Highways and Law and Justice.

For seven of these ministries, the DDGs contain ministry level object head wise expenditures directly. For three ministries, we had to sum up the expenditures across department DDGs. These include: the Ministry of Finance which is the sum of Economic affairs, Financial services, Revenue, Expenditure, and DIPAM. The Ministry of Health & Family Welfare is the sum of Health & Family Welfare, and Health Research. The Ministry of Home Affairs (MHA) covers Home Affairs, Cabinet, Police, and 8 departments for the UTs. The expenditure of Home Affairs, Cabinet and Police make up 74 percent of the MHA expenditure. In our analysis, we report the procurement expenditure for the MHA as the sum of these three departments, and do not include the UT departments in the analysis.

All rupee values used in these calculations are adjusted for inflation using the CPI index. The spending gaps are calculated for each year, from 2014-15 to 2018-19.

### Is there a link between consistent procurement and lower procurement under-spend?

We organise our findings into three figures, where each is in the form of a heat map. In these graphs, each row is a ministry, where the values going from left to right are those for 2014-2015 to 2018-2019. The deeper / stronger the colours, the higher the values. In all three figures, the order of the ministries stay the same: in descending order of estimated procurement expenditure in 2014-2015. The ministry with the highest procurement spending is on top.

In Figure 1, we present the estimated procurement spending of the 10 ministries. The Ministry of Road Transport & Highways had the highest estimated procurement spending (in Rs. billion) in 2014-2015, and the ministry of coal had the least. There is large variation in the procurement spend across the 10 ministries, as described in Sharma and Thomas (2021). Ministries such as Coal and Rural Development spend less than Rs. 1 billion on procurement and have consistently done so over last 5 years. Ministry of Road Transport & Highways is the top procuring ministry followed by Home Affairs, of which Department of Police is largest procurer. There is more variation across the years for the remaining ministries. For example, the ministry of Health & Family Welfare shows an increased estimated procurement expenditure in 2017-2018 compared to other years in this period. The Ministry of Law & Justice has a steadily increasing procurement expenditure during this period.

 Figure 1: Estimated procurement expenditure by ministry, 2014-15 to 2018-19

Figure 2 presents the estimated procurement spending gap from 2014-15 to 2019-20. Green stands for positive values (which is over-spending or under-budgeting). Red stands for negative values (which indicates under-spending or over-budgeting). Except for a few years, most ministries tend to under-spend. The exception is the Ministry of Road Transport & Highways, which is also the top procuring ministry.

 Figure 2: Estimated procurement spending gap (%) by ministry, 2014-15 to 2018-19

Figure 2 also shows a sharper amount of under-spending towards the latter part for some of the ministries. These include Finance, Civil Aviation, Rural Development and Coal. This contradicts the proposition that procurement capacity is systematically increasing over time.

These ministries also have a lower fraction of procurement in their spending. These ministries are less likely to develop procurement capacity. Among ministries with similar procurement spending patterns, the Ministry of Environment appears to be building procurement capacity, with a procurement spending gap closer to zero.

Lastly, Figure 3 presents the spending gap for total expenditure for these ministries across time. There is much more capacity among the ministries in managing to spend their overall budgets. Most of the heat-map shows colours mapping to values closer to 0. In fact, there is more evidence of over-spending than under-spending compared to Figure 2. The ministries of Finance and Rural Development have more instances of over-spending in their overall spending, while these are ministries which have estimated procurement under-spend.

 Figure 3: Total spending gap (%) by ministry, 2014-15 to 2018-19

### Alternative explanations

Underspend can be driven by other compulsions also. In a public choice theory worldview, state organisations would give the highest priority to wage and pension expenditures, and sacrifice procurement expenditures when faced with formal or informal budget constraints. This makes procurement budgets the most vulnerable to sudden cuts.

However, this argument should apply equally across all ministries. In fact, the ministries which do higher amounts of procurement should be the prime target for mid-year budget cuts. This argument predicts that under-spend should take place roughly everywhere, and maybe to a greater extent in the high-procurement ministries. The evidence, however, shows that under-spend is more prevalent in low-procurement ministries.

### Discussion

The main finding of this article is that ministries that tend to procure consistently, tend to have smaller procurement spending gaps. This is consistent with the idea that there is learning by doing, where doing procurement on a sustained basis gradually creates organisational capability for procurement.

If the key claim of this article is true -- that there is learning by doing in procurement -- this has a few interesting implications. In a department where there is low experience with procurement, the early years where procurement work begins will work poorly. In such a department, procurement under-spend is likely, with consequential failures in public expenditure programs and budgeting. In a department with low experience with procurement, a sudden jump in the procurement budget is likely to be associated with failure to spend. If the political leadership decides to push up the procurement spend of a department by three times, it would make sense to (a) increase the budget by only 20% and (b) initiate a capacity building program for procurement capability within that department.

Procurement is an expertise. No government organisation can sporadically do this well. It is an expertise which can be built, albeit over many years. In any government organisation, people and processes can be organised to focus on this expertise, and to devote time and effort on the entire pipeline of government contracting. The process of developing this capability can be accelerated by bringing in people with this specialised expertise. Strengthening the entire life cycle is required to successfully spend budget amounts. But this is only the beginning of success in procurement where government can contract to deliver quality projects efficiently, on time and at low cost.

### References

Aneesha Chitgupi and Susan Thomas. The make vs. buy decision of the union government, The Leap Blog. September 10, 2022.

Vijay Kelkar and Ajay Shah. In service of the republic: The art and science of public policy. Second edition, 2022.

Shubho Roy and Anjali Sharma. What ails public procurement: an analysis of tender modifications in the pre-award process, The Leap Blog, November 26, 2020.

Anjali Sharma and Susan Thomas. The footprint of union government procurement in India, XKDR Working Paper 10, November 2021.

Aneesha Chitgupi is a Research Fellow at XKDR Forum, Abhishek Gorsi is a doctoral candidate at the IGIDR and Susan Thomas is a Senior Research Fellowe at XKDR Forum and a Research Professor of Business at Jindal Global University. We thank Josh Felman, Sudha Krishnan, Ajay Shah and Anjali Sharma for feedback and comments.