## Tuesday, March 29, 2022

### How competitive is bidding in infrastructure public procurement? A study of road and water projects in five Indian states

by Charmi Mehta and Diya Uday.

### Introduction

Competition is central to the functioning of a market economy. Market power is a market failure, and governments around the world work hard to fight anti-competitive behaviour and market capture by firms. When competitive pressure is lacking, firms fail to achieve efficiency in production.

An efficient system for government procurement is one where the government obtains purchases for the lowest possible price. In the international literature, studies have shown the linkage between higher competition and lower procurement prices (Estache et al. 2008; Hanak and Muvchova 2015), greater efficiency (Adam et al. 2021) and a lower rate of corruption and kickbacks (Knack et al. 2017).

It is difficult to make normative claims about what is the adequate level of competition. Economists have emphasised contestability of a market as the underlying source of efficiency; simple proxies like concentration ratios do not correctly evoke the level of competition. Researchers in the field of competition levels in government contracting have used the number of bids received for a tender as an empirical measure.

There is some international evidence from developing countries about the desirable numbers of bidders in infrastructure public procurement. There are thumb rules, such as desiring eight-or-more bidders for a roads contract (Gupta 2002, Estache et al. 2008) or seven-or-more for water projects (Estache et al. 2008). These normative numerical values would of course, not readily carry forward across locales, but we use them cautiously in the present work.

In the field of government contracting in India, there is anecdotal evidence of anti-competitive behaviour in the market with reports of bid-rigging and collusion. In this article, we aim to step up from this to some statistical evidence. We ask: How much competition do we see in Indian infrastructure procurement? How does this vary across states and sub-sectors?

### Methodology

We hand-construct a novel data-set with a sample of tenders from five states: Tamil Nadu, Odisha, Maharashtra, Uttar Pradesh and Kerala. This choice of states was shaped by the levels of spending on infrastructure, geographical heterogeneity, and data availability. We extracted data from 1000 randomly sampled, awarded e-tenders published by state governments on the Central Public Procurement Portal (CPPP) in the water and roads sector for 2018 and 2019.

CPPP is a centralised repository of tender data at the union and state level. Procuring entities across union and state tiers are obliged to publish their tenders on the portal. We use data solely from this portal to ensure consistency in variables and recording. The data-set includes:

1. States in the sample: Kerala, Maharashtra, Odisha, Tamil Nadu, Uttar Pradesh
3. Years covered: 2018, 2019
4. Total sample size for each sector: 500 tenders
5. Key words used while searching through the CPPP to select tenders: "road", "water".

We extract the number of bids received for each tender in order to examine the level of competition.

In the research literature, it is argued that three factors shape competition in public procurement:

The value of the contract
Larger contract sizes require greater capital and expertise, which act as an entry barrier for small/mid-size players in the market (Estache et al. 2009; McEvoy 2020).
Structure of the tender
Bundled tenders, where multiple works of different types are bundled into a single tender document, restrict competition to only those firms that can undertake the varied components of bundled tender (Estache et al. 2009). Dividing contracts into smaller lots bolsters MSME participation and thus competition (Hoekman et al. 2022).
Time taken to award contracts
Lengthy award schedules require participating firms to lock-in capital and resources for the bid amidst the uncertainty of winning the bid; this adversely impacts private sector enthusiasm towards bidding (World Bank 2020).

We will examine the extent to which the number of bids per tender correlates with these features.

### Results

Table 1 summarises the statistics on the number of bids received across the five sample states for the years 2018 and 2019.

Table 1: Number of bids received in the infrastructure sector

2018 2019

Min Max Median Average Min Max Median Average
Maharashtra 1 4 3 3.26 1 11 3 3.94
Uttar Pradesh 1 11 3 4 2 9 3 3
Tamil Nadu 1 3 2 2 1 3 2 2.08
Kerala 1 8 2 2.08 1 7 2 1.80
Odisha 1 27 5 7 1 55 6 9.96

In the Water sector

2018 2019

Min Max Median Average Min Max Median Average
Maharashtra 1 34 3 5.56 1 17 3 4.17
Uttar Pradesh 1 8 3 3.04 1 16 3 4.05
Tamil Nadu 1 4 2 1 1 12 3 2
Kerala 1 5 2 2.08 1 5 3 2.28
Odisha 1 52 2 2.45 1 27 3 4
Source: Authors' compilation and calculation from CPPP data

Q.1. How competitive is infrastructure procurement?

We find that, with the exception of the road sector in Odisha, the level of competition in terms of the average number of firms bidding for the projects is lower than the normative thumb-rules from the literature. This holds across all states, sectors and years. Further, two features about the lack of competition in this data-set merits discussion:

1. Tenders where bids satisfy the thumb-rules of the literature

We examine the fraction of tenders in our sample that satisfy these thumb rules, and count the tenders that got more than seven bids in the water sector and eight bids in the roads sector. Table 2 summarises these results. Here, we see that the roads sector fares worse than the water sector.

Table 2: Fraction of tenders that receive bids meeting normative thumb rules (share in per cent of tenders)

Roads 5 5 0 34 1
Water 14 7 1 8 0

2. Tenders that received only one bid and were awarded

We find several awarded tenders that attracted only single bids. This is true even in states with relatively higher levels of competition such as Uttar Pradesh, Maharashtra and Odisha. But there are more tenders with single bid awards in the states with the lowest levels of competition, namely, Tamil Nadu and Kerala. We find that single bid awards are more prevalent in the water sector.

Table 3: Fraction of awarded tenders that received only one bid (% of tenders in the sample)

Roads 6 2 4 6 40
Water 8 6 52 8 21

Q.2. How do the findings vary across states and sectors?

The results are fairly consistent results across sectors and years. For instance, three states -- Uttar Pradesh, Maharashtra, and Odisha -- have higher levels of competition across both sectors when compared to the other observed states. Kerala and Tamil Nadu have lower levels of competition across both years for both sectors.

Q.3. What features of tenders correlate with a low number of bids?

In the literature, there has been interest in three features that may shape the level of competition: the value of the contract, the structure of the tender and the time taken to award contracts. In our data-set, however, these three factors do not correlate with the number of bidders.

### Discussion

This evidence suggests there is a low level of competition in public procurement, in two sectors and five states. There are some systematic patterns, where some states and sectors fare worse in getting competitive bidding than others.

Competitive conditions seem to be the feature of a given state. This suggests that there are some features in states like Kerala or Tamil Nadu, which are inhibiting competition, and can be addressed in a way that would impact on government purchases across sectors.

A large number of tenders with a single bid that get awarded are a curious phenomenon and merit further research. These tenders are awarded under the previous Central Vigilance Commission (CVC) guidelines, which require that state Public Works Departments (PWD) cancel tenders that receive single bids at the first instance. Single bids could be accepted only if the procuring entity received only one bid, even after re-tendering. However, the recent General Instructions on Procurement and Project Management, allows the acceptance of single bids under certain conditions. These include: (i) the procurement was satisfactorily advertised and sufficient time was given for bid submission; (ii) the qualification criteria was not unduly restrictive; and (iii) the price in the bid is reasonable in comparison to market values.

Further research is required in extending this kind of work to other sectors and locales, to assess the extent to which the lack of competition is a more general phenomenon in public procurement in India. The source of this lack of competition also merit exploration.

There are limitations in how state organisations do procurement (Mehta and Thomas 2021), including potential gaps in the capacity of implementing rules (Roy and Uday 2020), inefficiencies of processes and timelines (Roy and Sharma 2020), and delayed payment of invoices (Mannivanan and Zaveri 2021). Such problems could create an inhospitable environment for bidding firms, and deter many good firms from taking interest in state purchases. Well incentivised state actors should solve these problems. This raises questions about the feedback loops that impinge upon state actors.

### References

Antonio Estache and Atsushi Iimi, (Un)bundling Infrastructure Procurement: Evidence from Water Supply and Sewage Projects, Policy Research Working Paper No. 4854, World Bank, March 2009.

Antonio Estache and Atsushi Iimi, Procurement Efficiency for Infrastructure Development and Financial Needs Reassessed, Policy Research Working Paper No. 4662, World Bank, March 2008.

Bernard Hoekman and Bedri Taş, Policy and SME participation in public procurement, Vox EU - CEPR, 23 March 2022.

Charmi Mehta and Susan Thomas, Lessons from the COVID-19 vaccine procurement of 2021, The LEAP Blog, 15 November 2021.

Emma McEvoy, Small and Medium-Sized Enterprises (SME) Participation in Public Procurement, Maynooth University, 2020.

Isabelle Adam , Alfredo Hernandez Sanchez and Mihály Fazekas, Global Public Procurement Open Competition Index, Government Transparency Institute, Working Paper Series: GTI-WP/2021:02, April 2021.

Pavithra Mannivanan and Bhargavi Zaveri, How large is the payment delays problem in Indian public procurement?, The LEAP Blog, 22 March 2021.

P. Manoj, Govt lifts the ‘fear’ on accepting single bids during public procurement tenders, The Hindu, 2 November 2021.

Srabana Gupta, Competition and collusion in a government procurement auction market, Atlantic Economic Journal 30, 13–25, 2002.

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

Shubho Roy and Diya Uday, Does India need a public procurement law?, The LEAP Blog, 19 August 2020.

Stephen Knack, Nataliya Biletska and Kanishka Kacker, Deterring Kickbacks and Encouraging Entry in Public Procurement Markets : Evidence from Firm Surveys in 88 Developing Countries, World Bank Working Paper, May 2017.

Tomáš Hanák and Petra Muchová, Impact of Competition on Prices in Public Sector Procurement, Procedia Computer Science, Volume 64, Pages 729-735, 2009.

World Bank, Contracting with the government, World Bank Doing Business, 2020

Charmi Mehta and Diya Uday are CMI-XKDR Forum researchers. The authors thank Shailesh Phatak, Susan Thomas and Ajay Shah for their valuable inputs on this work; and Abhinav M from the Indian Institute of Human Settlements for his valuable research support.

LaTeX mathematics works. This means that if you want to say $10 you have to say \$10.