d4

SANDIP SUKHTANKAR*
Dartmouth College
MILAN VAISHNAV†
Carnegie Endowment for International Peace
Corruption in India: Bridging Research

Evidence and Policy Options§
ABSTRACT Corruption has become an increasingly salient issue in India today,

spawning enormous interest from the media as well as a large amount of academic
research. Yet, there is a sizeable gap between what has captured the media’s
attention, the policy options under discussion, and the actual evidence base drawn
from empirical research on corruption. We attempt to bridge this gap by directly
addressing the particular challenges that corruption in India poses. Academic evi-
dence supports the popular perception that corruption is widespread and endemic.
However, we find that the costs of day-to-day corruption are just as large, if not
larger, than those of the “scams” that dominate headlines. Further, we find that there
is very little evidence to support the idea that greater transparency, information,
and community-based efforts have a significant impact on reducing corruption on
their own. This is also true for some technological interventions, although those
interventions—like direct benefit transfers—that bypass middlemen and corrupt
officials have a much greater scope for success, as do interventions that transfer
bargaining power to citizens and beneficiaries. We find much to commend in the
sensible and wide-ranging legislative agenda to combat corruption, including the
Right to Service and Public Procurement bills. However, what is most important
for combating corruption is not the law on paper but the implementation of the law;
the binding constraint, as always, is the government’s desire and ability to punish
corrupt officials and politicians.
Keywords: Corruption, India, Transparency, Public Sector, Political Financing,

Public Sector Recruitment, Electoral Reform
JEL Classification: D4, D73, H10, H40, H83, K42, O10
*sandip.sukhtankar@dartmouth.edu
†mvaishnav@ceip.org
§We thank our discussants Devesh Kapur and Karthik Muralidharan, the Editor, Shekhar
Shah, and Pranab Bardhan, Vijay Joshi, and other participants at the 2014 India Policy Forum
for comments. William Hayes and Danielle Smogard provided excellent research assistance.
Milan Vaishnav would like to acknowledge financial support from the Smith Richardson
Foundation. All errors are our own.
193
194 I N D I A P O L I C Y F O R U M , 2014–15
1. Introduction
C orruption in India is a topic that seems to never fall out of fashion.

From as far back as Kautilya’s Arthashastra in the 4th century BC
to the 2G telecommunications spectrum scam in the contemporary period,
corruption is widely perceived to be an endemic phenomenon in the Indian
subcontinent.
Yet, by any measure, the salience of corruption in the public policy dis-
course in India has ratcheted up in recent years. This is, in part, a reflection
of a series of high-profile “scams” that plagued the recently departed United
Progressive Alliance (UPA) government. In 2011, India saw a groundswell
of popular protest in which thousands of citizens joined in anticorruption
demonstrations after a series of scandals implicated ruling politicians and
their cronies in billions of dollars of graft—from the Commonwealth Games
to 2G scandals, and from “Coalgate” to Adarsh Housing Society. This sim-
mering discontent later gave rise to a new political party—the Aam Aadmi
Party (AAP)—that burst onto the political scene with a pledge to clean up
government.
While the AAP’s popularity has largely been limited to Delhi and its
surrounding areas, the recent anticorruption mood in India arguably helped
propel the Bharatiya Janata Party’s (BJP) Narendra Modi and his National
Democratic Alliance (NDA) government into power in the 2014 general
election. Indeed, Modi consistently invoked his fight against corruption on
the campaign trail, telling huge crowds that the Congress Party-led UPA
government stood for the “ABCD of corruption,” listing numerous scams
in which the party and family members of the Nehru–Gandhi dynasty were
implicated: “A for Adarsh, B for Bofors, C for Commonwealth Games, and
D for Damad Ka Karobaar (‘son-in-law’s business,’ a reference to corrup-
tion allegations against Robert Vadra, the son-in-law of Congress president
Sonia Gandhi).”
A post-election analysis conducted by the Centre for the Study of
Developing Societies (CSDS) suggests that anticorruption sentiment was
a key contributor to the BJP’s winning an outright majority in parliament,
the first time any party has done so in three decades (and the first time in
history such a feat was accomplished by a party other than the Congress).
According to CSDS’ 2014 National Election Study, only concerns over
inflation and lack of economic development were more important than cor-
ruption in determining voters’ choices in the election.
In this paper, we try to bridge the gap between evidence and policy when
it comes to understanding the causes and consequences of corruption in
Sandip Sukhtankar and Milan Vaishnav 195
India and formulating solutions to address its spread. This gap exists for
several reasons.
For starters, corruption is by its very nature difficult to objectively
measure. Most corrupt transactions transpire out of public view and the
parties involved have incentives to keep it that way. What emerges from
media reporting is, by definition, ex post and often sensationalist in nature.
For example, the media focuses on “scams” that involve billions of rupees
worth of malfeasance; yet our calculations suggest that the costs of day-to-
day corruption are at least of the same order of magnitude, if not higher. We
compiled an inventory of the biggest public corruption scandals uncovered
after the year 2000,1 finding that the amounts involved sum up to hundreds
of billions of dollars (the mean scam “value” was `36,000 crore, and the
median `12,000 crore; see Table 1).2 Eye-popping as these numbers are,
the costs of day-to-day corruption are comparable: e.g., Muralidharan et al.
(2014b) calculate the annual costs of teacher absence to be in the range of
`8,100–9,300 crore; Transparency International and CMS (2005) estimate
the costs of bribes paid annually for accessing various government services
across India to be `21,000 crore. Hence, attention-grabbing one-off scandals
may deflect attention from corruption that is just as costly, but is harder to
find and measure.
Second, while the existing social science literature has established some
theories of corruption, these have produced markedly divergent predictions
about both the causes and consequences of corruption. In particular, existing
theory is ambiguous about whether corruption is bad for the economy. An
old literature suggests that corruption “greases the wheels” of the economy
by providing incentives for bureaucrats to work harder, and also by allow-
ing firms and individuals to get around costly and inefficient red-tape and
regulations (Huntington 1968, Leff 1964). Another strand of thought predicts
that corruption may have no efficiency effects, only redistributive ones;
e.g., if the most efficient firm is the one that can pay the highest bribes to
1. While there is no precise formula for determining what constitutes a big scandal, we
began by scanning lists of “corruption scams” compiled by news outlets (India Today, Outlook,
NitiCentral, Yahoo, DNA) over the past several years. We then created a shortlist so that
only scams that featured on multiple lists were included. For some number of scams, it was
difficult to find the level of detail we wanted, so we excluded these. Table 1 summarizes the
28 scandals we examine.
2. Not all of these amounts involve losses to the government and, moreover, these amounts
must be viewed with caution, since the media tends to inflate and focus on the largest numbers.
Indeed, the “costs” written up in the media conflate the value of bribes that changed hands, pure
theft or embezzlement from the government of various sorts, cheating the exchequer out of the
appropriate value of assets, as well as private losses in which one party simply cons the other.
TABLE 1. Major Corruption Scandals in India in the 21st Century
Name Date Years Sector State Cost (`Crore) Cost type
Taj Heritage Corridor Scam 2002–03 2 Construction Uttar Pradesh 175 Embezzlement
Uttar Pradesh NRHM Scam 2005–11 7 Construction Uttar Pradesh 10,000 Embezzlement
Tatra Trucks Scam 1997–11 15 Defense N/A 750 Bribes
Agusta Westland Chopper Deal Scam 2010–13 4 Defense N/A 450 Bribes
Calcutta Stock Market Scam 2001 1 Financial West Bengal 120 Private fraud
Telgi Stamp Scam 2003–13 11 Financial Maharashtra, others 43,000 Private fraud
IPO Demat Scam 2003–05 3 Financial N/A 146 Private fraud
Saradha Group Chit Fund Scam 2006–13 8 Financial West Bengal, others 20,000 Private fraud
Sahara India Pariwar - Investor Fraud Case 2008–10 3 Financial Uttar Pradesh 24,000 Private fraud
Kerala Solar Panel Scam 2010–13 4 Financial Kerala 7 Private fraud
Rice Export Scam 2008–09 2 Food grains N/A 2,500 Value loss to govt.
Uttar Pradesh Food Grain Scam 2002–10 9 Food grains Uttar Pradesh 35,000 Embezzlement
Gegong Apang Public Distribution System Scam 1995–04 10 Food grains Arunachal Pradesh 1,000 Embezzlement
Antrix Devas/ISRO Spectrum Allocation Scam 2005–11 7 IT N/A 2,00,000 Value loss to govt.
2G Spectrum Scam 2008 1 IT N/A 56,000 Value loss to govt.
Satyam Computer Services Scandal 2009 1 IT N/A 14,162 Private fraud
Karnataka Wakf Board Scam 1954–11 58 Land Karnataka 2,00,000 Value loss to govt.
Maharastra Adarsh Housing Society Scam 2003–10 8 Land Maharastra 163 Value loss to govt.
Andhra Pradesh Land Scam 2006–12 7 Land Andhra Pradesh 1784 Value loss to govt.
Noida Corporation Farm Land Scandal 2009–11 3 Land Uttar Pradesh 5,000 Value loss to govt.
Maharashtra Irrigation Scam 1999–09 11 Land Maharashtra 35,000 Embezzlement
Odisha Mine Scam 2000–10 11 Mining Odisha 50,000 Value loss to govt.
Coalgate 2004–12 9 Mining N/A 1,86,000 Value loss to govt.
Bellary Mining Scandal 2005–11 7 Mining Karnataka, 21,000 Value loss to govt.
Andhra Pradesh
Jharkhand Mining Scam 2006–09 4 Mining Jharkhand 3,400 Value loss to govt.
Goa Mining Scam 2009–11 2 Mining Goa 35,000 Value loss to govt.
Cash for Votes Scandal 2008 1 Political N/A 50 Bribes
CWG Scam 2010 1 Procurement N/A 70,000 Embezzlement
Source: Authors' compilation from various media sources. Full details provided in Appendix A in paper on first author's website, http://www.dartmouth.edu/~sandip/Sukhtankar-
Vaishnav-Corruption-IPF_Full.pdf
officials in order to obtain contracts or licenses from the government, then

there is no efficiency consequence, just a transfer from the government to
the corrupt official (Lui 1985).
On the other hand, there are a number of theoretical reasons why cor-
ruption might negatively affect efficiency and economic and political
development. Continuing the example above, the secrecy inherent in cor-
ruption might mean that firms connected to the bureaucrat, which might not
necessarily be the most efficient ones, obtain contracts or licenses (Shleifer
and Vishny 1993). Meanwhile, not all rules and regulations are inefficient,
and in cases where individual willingness or ability to pay diverges from
what is considered socially good, corruption will not be optimal (Banerjee
1997). For example, being able to drive a car is a reasonable requirement for
obtaining a driver’s license, and bribing to get around this rule might reduce
social welfare. Further, bribery might increase bureaucrats’ incentives to
create inefficient red-tape in the first place (Banerjee 1997).
Third, the industrial organization of corruption may also matter. Suppose
a firm needs multiple clearances to set up a project. If a number of decen-
tralized corrupt agents act as independent monopolists, they will charge
bribes that are “too high” and create an inefficient entry barrier (Shleifer
and Vishny 1993). Additionally, corruption that involves straight theft might
distort optimal public finance (Niehaus and Sukhtankar 2013b). Finally, the
presence of widespread corruption in the economy might incentivize rent-
seeking rather than productive activities (Murphy et al. 1991).
Fortunately, one recent bright spot from the research community is the
development of a burgeoning empirical literature on corruption in India that
tests many of these theoretical predictions. These studies have produced a vast
amount of knowledge about both the political economy of corruption as well
as the relative effectiveness of various solutions in addressing this scourge.
However, much of this scholarly work has not filtered down into the
policy domain. While there are several excellent recent reviews of research
on corruption drawing on a wide array of settings (Banerjee et al. 2012,
Olken and Pande 2012, Pande 2007), corruption in India poses particu-
lar challenges that these surveys do not explicitly address. For example,
India’s archaic campaign finance laws result in candidates turning to illicit
means to raise funds for elections (Kapur and Vaishnav 2015; Sukhtankar
2012). Moreover, electoral accountability mechanisms proven to check
corruption in other contexts (Ferraz and Finan 2011) fail in India where
criminal and corrupt politicians thrive in spite of these measures (Aidt
et al. 2013, Banerjee and Pande 2009, Vaishnav 2012). Furthermore, many
such reviews do not produce explicit recommendations for formulating

better public policy.
These various misalignments have created a great deal of confusion.
Despite the increasing salience of corruption in India and the heated political
rhetoric the subject arouses, there is a large gap between what has captured
the media’s attention, the policy options under discussion, and the actual
evidence base from empirical research on corruption. To give one example,
the conventional wisdom holds that corruption in politics or in public works
programs is often the result of information asymmetries. Yet, multiple stud-
ies actually show that information provision is largely ineffective in pro-
ducing better governance outcomes, at least in isolation (Banerjee, Banerji,
Duflo, Glennerster, and Khemani 2010a; Niehaus and Sukhtankar 2013b;
Ravallion, van de Walle, Dutta, and Murgai 2013).
The objective of this paper is to make a modest contribution toward a
more optimal alignment. To begin with, we define corruption using the
most common academic definition: “the misuse of public office for private
gain” (Bardhan 1997). Note that this immediately narrows our focus to
the public sector, to include both bureaucrats and politicians. Our choice
of definition does not imply that we are leaving out the private sector or
private individuals, usually the source of bribes that represents the “private
gain” of the bribe receivers; it simply means we are ignoring purely private
fraud, e.g., a case of a private firm paying off another private firm to get an
unfair advantage in a private transaction between them.3 Given the paucity
of research in this area and the difficulty in conceptualizing every type of
private sector malfeasance, we focus on public sector corruption.
Having defined corruption in this manner, we divide the remainder of the
paper into four main parts. In Section 2, we provide a stylized classification
of the causes of corruption in India. We identify four underlying drivers:
two “deep” causes (lack of enforcement capacity, regulatory complexity)
and two “proximate” causes (inadequate regulation of political finance,
shortcomings in public sector recruitment and postings).4
In Section 3, we pivot from a discussion of corruption’s underlying
drivers to presenting a rubric for classifying corrupt acts. Specifically, we
organize corruption into three general categories based on the actions of
public officials: facilitative, collusive, and extractive corruption. Using
3. However, some of the scams that we identify in Table 1 involve financial sector fraud
with only private actors.
4. Technically speaking, the issues of political finance and discretionary transfers and
postings are both drivers of corruption, as well as corrupt outcomes in and of themselves. In
the latter regard, they are manifestations of collusive and/or extractive corruption.
this simple framework, we then draw on research from both economics and
political science to describe the magnitudes, causes, and consequences of
each of these types of corruption.
Section 4 discusses broad strategies for combating corruption, describes
major recent anticorruption legislation either passed or under discussion, and
explores academic evidence that evaluates the effectiveness of both broad
strategies and particular legislation in combating corruption. In Section 5,
we discuss India’s unique political economy of reform and what, if anything,
India can learn from the historical record. In Section 6, we conclude with
some parting lessons.
It is important to note that our focus is on academic research that has
rigorously evaluated causal relationships with the best quality data and infor-
mation possible. Given the vastness of the literature on corruption in India,
we were compelled to narrow our parameters in this way. Furthermore,
with an eye towards distilling the major takeaways of this literature and
extracting the core policy prescriptions that emerge, we have generally tried
to avoid discussions of empirical methodology. In so doing, we follow a
template similar to that pursued by Muralidharan (2013) in his review of
education in India.
2. Causes of Corruption
In this section, we briefly review the major causes of corruption in India.

Given the complexity and breadth of an issue such as corruption, it is impos-
sible to fully account for all of the underlying drivers that create incentives
for corrupt behavior. Rather than attempting to construct such an unwieldy
inventory, we focus instead on stylized drivers of corruption.
India’s per capita Gross Domestic Product (GDP) in 2015 places it among
the ranks of the world’s lower middle-income countries. Given the inverted-
U shaped relationship between corruption and per capita GDP across
countries, it seems intuitive that corruption in India might pose a particular
challenge. In the poorest countries, corruption is limited because there is,
frankly, not much to steal; in the most undeveloped economies, there is
little need to delegate the types of discretionary tasks that lend themselves to
bribery. On the opposite end of the spectrum, the rich, advanced industrial-
ized countries have developed robust legal frameworks and institutional
and enforcement mechanisms necessary to deter and combat corruption.
As Laffont (2006) points out, one is likely to find the greatest corruption in
those countries standing at the unfortunate juncture of a heavily regulated

economy that does not yet boast adequate enforcement capacity.
We refer to these two factors—lack of enforcement capacity and regula-
tory complexity—as “deep” causes, insofar as they reflect the key institu-
tional parameters that define India’s corruption environment. The remaining
two drivers— inadequate regulation of political finance and shortcomings
in public sector recruitment and postings—are more “proximate” in nature.
By proximate, we do not mean to imply that they can be easily addressed
in the short-term, but rather that they are well-defined pathologies spawned
by larger infirmities plaguing India’s institutional moorings. Based on the
literature, we believe these four drivers set the stage for the vast majority
of corrupt acts taking place in India in recent years.
2.1. Regulatory Complexity

In light of India’s experience of colonial exploitation and the realities of
its abject poverty, and in concert with prevailing ideological trends of the
era, India’s first Prime Minister, Jawaharlal Nehru, launched India down
a path of state-led development. Upon Nehru’s death, his daughter Indira
Gandhi doubled down on this state-led model, triggering the onerous system
of state controls later dubbed the “License Raj.” With the “pro-business”
reforms of the 1980s and the “pro-market” reforms of the early 1990s, the
role of the state in India’s economy appreciably diminished (Rodrik and
Subramanian 2005). However, more than two decades after enacting liberal
economic reforms that loosened controls on private business and deepened
India’s economic integration with the rest of the world, many vestiges of
the License Raj persist and help facilitate significant rent-seeking activity.
First, India remains an intensely difficult place for firms to do business.
According to the 2015 World Bank Doing Business indicators, India ranks
142 out of 189 countries. Its standing compares unfavorably to its South
Asian peers, as well as other “lower middle income” economies, and even
within the universe of BRICS nations. For instance, according to the World
Bank survey, the act of obtaining a single construction permit in India
involves 27 discrete procedures, takes 162 days, and costs 46 percent of the
total outlay to a construction firm of building a warehouse.
Second, lucrative sectors of the economy remain largely untouched by
economic reforms, allowing the state to exercise a heavy hand through its
regulatory authorities as well as the market power of public sector undertak-
ings. As Reserve Bank of India Governor Raghuram Rajan has noted, the
state in India continues to dominate the “commanding heights,” such as oil
and gas, mining, and heavy industry (Rajan 2012). The possibilities of quid
pro quos are especially high in these areas, as the allocation of rights over
natural resources has been poorly defined historically and virtually all but
ignored by successive economic reforms.
Two facts result from this state of affairs. First, the regulatory intensity
of the state with respect to private business activity not only minimizes the
role for market forces, but also facilitates a natural quid pro quo whereby
the state provides licenses, permissions, clearances, etc. in exchange for side
payments. This is why India reliably rates poorly on most perception-based
indicators of corruption and bribery.
Second, corruption is most intense in those sectors where the regulatory
footprint of the state is the greatest. Consider, for instance, the sectoral
breakdown of our inventory of scams listed in Table 1. The clustering
of scams within certain sectors is quite instructive; the mining and land
sectors are most commonly represented, accounting for 35 percent of the
total (10 of 28). This category includes scams such as the now infamous
“Coalgate” scandal, in which the Comptroller and Auditor General of India
(CAG) accused the Union Government of allocating almost 200 coal blocks
without a competitive bidding process between 2004 and 2009. The CAG
alleged that private firms paid far less than what they would have had the
licenses been auctioned. Many of these firms were owned by, or closely
linked with, sitting politicians, implying that political criteria were used to
allocate licenses.
As argued above, this interaction of regulatory intensity and corruption
is in line with our priors, as these areas represent obvious sources of rents
in the economic sense of the word—they are sectors of the economy where
the regulatory intensity of the state is immense and the opportunities for
bribes or kickbacks are legion. The values involved are also generally high
in these sectors, given their centrality to the broader economy.
2.2. Lack of Enforcement and Implementation Capacity

While regulatory complexity allows wide scope for rent-seeking and extrac-
tion, the capacity to combat misdemeanor and enforce rules and regulations
is seriously limited. The government agencies in charge of administration
and law and order are overburdened, inadequately staffed, and often poorly
equipped. Thus implementing complex rules and policies, as well as catching
and punishing rule-breakers, is a massive challenge.
Given the discussion on the cumbersome role of the state in the previous
section, it is somewhat counterintuitive that India has the smallest number of
government employees as a ratio of its population among any of the Group
of 20 (G-20) nations, with only 146 public sector employees per 10,000
residents (Figure 1). Compared to China (537), the USA and Germany (both
about 730), or Russia (1534), this number is remarkably low. Even this low
number is likely an overestimate of administrative capacity, given that it
includes employment in the state-owned Indian Railways, one of the larg-
est employers in the world. Furthermore, the overall strength of the public
sector has declined since 2001, although this partially reflects a divestment
from public sector enterprises and a move to greater contract employment,
rather than a reduction in administrative capacity per se (Figure 2).
Nevertheless, what the relatively small number of public employees
means for a country India’s size is that the administrative service has an
enormous burden placed on it to not only implement the plethora of schemes
and regulations in place but also to enforce them. For example, the chief
administrative officer of the most relevant administrative unit for imple-
menting public programs in India—the district collector—is also the official
in charge of law and order in a district, which has on average 2 million
inhabitants. While these officers are selected from the cream of the crop of
the Indian Administrative Service (IAS), given the enormous burden placed
on their shoulders, the decks are stacked against them from the beginning.
FIGURE 1. Public Sector Employment in G-20 Countries

1600
Public employees per 10,000 residents
1400
1200
1000
800
600
400
200
0
Russian Federation
Canada
United Kingdom
France
China
Australia
Germany
United States
Italy
Brazil
Argentina
Japan
Indonesia
Turkey
Mexico
Saudi Arabia
South Africa
India
Sources: International Labor Organization; Saudi Arabia Ministry of Economy and Planning; China National
Bureau of Statistics; World Bank.
FIGURE 2. Public Employment in India from 1971–2001 (millions)

25
20 19.1 19.1
17.5
Public employment (million)
15.5
15
10.7
10
0
1971 1981 1991 2001 2011
Year
Source: Economic Survey of India (various years).
Compounding the administrative burden they face, officers regularly face

political interference. Although rules were designed to prevent meddling,
Krishnan and Somanathan (forthcoming, p. 11) write that, “the power to pun-
ish arbitrarily has been acquired (and used to telling effect) by the political
executive … through misuse of the power of transfer.” Frequent transfers,
or the very threat of transfers, make officers’ jobs even harder, as they only
have “short tenures in each post, which greatly diminishes effectiveness.”
The state of public administration is not helped by the fact that many
government jobs are left unfilled, possibly due to political jockeying (see
Section 2.4 below). The problem of unfilled job vacancies is particularly
severe for the police and courts, the pillars of the enforcement arm of the
government. To begin with, the number of police and judicial officials as
a proportion of the population is very low. By our rough estimate, India
has approximately 16.5 judges per one million residents, which compares
unfavorably to around 101 judges for a comparable population in the USA.
In addition, India has the lowest rate of police officers per capita—122.5 per
100,000 people—of any G-20 member state. Yet even with this low base, a
quarter of police vacancies across the country are unfilled. For certain states,
the vacancy rate is alarming: Uttar Pradesh, which faces serious law and
order difficulties, has nearly 60 percent of its police posts unfilled (Table 2).
TABLE 2. Police Vacany Rates across Indian States, 2013

State Sanctioned Actual Vacancy Rate (%)
Andhra Pradesh 107730 89787 16.66
Arunachal Pradesh 5322 4548 14.54
Assam 30979 19870 35.86
Bihar 78641 62803 20.14
Chhattisgarh 54499 32259 40.81
Goa 5330 4722 11.41
Gujarat 75267 54921 27.03
Haryana 48569 39059 19.58
Himachal Pradesh 10833 9864 8.94
Jammu & Kashmir 51860 47755 7.92
Jharkhand 55771 42073 24.56
Karnataka 82358 68584 16.72
Kerala 40185 37995 5.45
Madhya Pradesh 76518 62281 18.61
Maharashtra 191197 182832 4.38
Manipur 18636 12219 34.43
Meghalaya 7490 5948 20.59
Mizoram 4321 3377 21.85
Nagaland 6013 5637 6.25
Odisha 35047 30850 11.98
Punjab 59201 59969 –1.30
Rajasthan 88881 69689 21.59
Sikkim 2381 1906 19.95
Tamil Nadu 99541 87581 12.02
Tripura 12898 10441 19.05
Uttar Pradesh 331984 133261 59.86
Uttarakhand 17074 16643 2.52
West Bengal 101560 72083 29.02
A&N Islands 3741 3453 7.70
Chandigarh 5794 5414 6.56
D&N Haveli 354 287 18.93
Daman & Diu 422 370 12.32
Delhi 72686 67686 6.88
Lakshadweep 606 384 36.63
Puducherry 2423 2433 –0.41
Total (All) 1786112 1348984 24.47
Source: National Crime Records Bureau, Crime in India, 2013.
The judiciary faces the same dilemma: almost 30 percent of seats on

High Courts across the nation are vacant, in addition to 22 percent of
seats in district and subordinate courts. Only the Supreme Court seems to
be relatively free of this malaise, as the corresponding rate there is only
6.5 percent (Figure 3).
One of the main consequences of these judicial vacancies is that it takes
an inordinately long time to resolve cases. This fact should come as no
FIGURE 3. Unfilled Vacancies in Judicial System

35%
29.4%
Percent of Sanctioned Seats Unfilled
30%
25% 22.3%
20%
15%
10%
6.5%
5%
0%
Supreme Court High Courts District/Subordinate Courts
Source: Supreme Court of India, Court News (various years).
Note: Measured at 12/31/2013 for Supreme and High Courts, 9/30/2013 for District/Subordinate.
surprise to anyone in India, but the figures do not make for enjoyable read-
ing. As of 2011, approximately 24 percent of court cases had been pending
for at least five years, while 9 percent had been pending for more than 10
years (Law Commission of India 2014). At the start of 2014, there were a
total of 31.4 million cases pending across all courts in India (Figure 4). While
the situation is slowly improving—for all courts the number of outstanding
cases on January 1, 2013 was higher than the end of year figures—the sheer
level of backlog induces despair.
We discuss initiatives to tackle these administrative and judicial issues
in the final section. But as of now, we merely stipulate the fact that India’s
capacity to enforce rules and regulations in virtually every domain is
severely wanting.
2.3. Inadequate Regulation of Political Finance

Turning now to proximate drivers of corruption, India’s approach to politi-
cal finance is emblematic of these larger weaknesses of enforcement and
regulation. What makes this failing puzzling is that it stands in sharp contrast
to the actual planning and execution of elections, which are ably handled
by the independent Election Commission of India (ECI), one of the most
autonomous election agencies in the world. At the heart of this puzzle sits
the aforementioned overbearing role of the state in the economy. As a recent
analysis of India’s political finance regime concludes, “Until the Indian
state retreats from major sectors of the economy and gives way to market
FIGURE 4. Outstanding Cases in Indian Courts, 2013

3,00,000 Opening Case Balance January 2013
Pending Cases December 2013
2,50,000
2,00,000
1,50,000
1,00,000
50,000
0
Supreme High (100s) District/Subordinate (100s)
Source: Supreme Court of India, Court News (various years).
forces, politicians and business will have reason to perpetuate a system of

trading policy and regulatory favors for monetary payments and campaign
‘donations’” (Sridharan and Vaishnav, forthcoming).
Buttressing this system of favor trading are shortcomings in the underly-
ing legal and regulatory powers of the ECI to adequately regulate political
finance. For instance, corporations and parties are only legally required
to publicly disclose political contributions in excess of `20,000. This rule
allows contributors to package unlimited political contributions just below
this threshold value completely free of disclosure. Indeed, in 2014 the
Association for Democratic Reforms (ADR) reported that 75 percent of
the income of India’s six major parties comes from undocumented sources
(ADR 2014).
On the expenditure side, candidates face strict limits on spending once
elections have been announced, but election authorities struggle to prop-
erly verify their reported expenditure since a substantial portion typically
occurs “in the black.” Even though the ECI has devoted greater resources to
addressing this problem, it faces several challenges. Under existing statute,
the ECI lacks clear powers to take follow-up action in the event a candidate
files false or misleading declarations. An even bigger problem lies with a
loophole in the law that allows candidates to keep secret party and sup-
porter expenditures on behalf of their campaigns that are spent propagating
the party program rather than endorsing the specific candidate in question
(Sridharan and Vaishnav, forthcoming).
In theory, if parties were transparent in their finances, some triangulation

would be possible. Unfortunately, party finances are highly opaque. Parties
are required to submit audited accounts to the ECI, but there is no require-
ment that the auditor needs to be an independent, third-party entity, which
creates incentives for parties to cook the books. The ECI recently issued
guidelines to parties to strengthen the reporting of their finances, but without
additional legal authorities, the ECI has no credible sanctioning mechanism.
In light of these infirmities, the regulation of political finance in India
is in dire shape and, hence, susceptible to corruption. Of the themes high-
lighted in this paper, political finance is arguably in greatest need of further
research and exploration. Commentators across the political spectrum have
recognized its centrality in corruption dynamics in the country. For instance,
Mehta (2002) has noted: “The reform, regulation and overhaul of the means
by which political parties and candidates finance elections is arguably the
single most important institutional challenge facing Indian democracy.”
The Economist (2014) summarized the issue in the context of the recently
completed 2014 general election more poetically: “picture the elections as
a dark sea of liquid assets, mostly undocumented cash (and a lot of liquor
too), overspilling the dykes that were meant to keep it in check.”
The opacity of political finance, however, presents an obvious obstacle
to careful empirical work. Hence, there is space for creative “forensic”
approaches.5 One such study is Sukhtankar (2012). The author finds evi-
dence of electoral cycles in input prices paid for sugarcane among politically
controlled mills in Maharashtra. Specifically, he finds that cane prices paid
to farmers by politically controlled mills falls in election years. Sukhtankar
claims that sharp drops in cane prices represent mill funds siphoned off to
finance politicians’ electoral campaigns; in other words, these funds serve
as indirect political contributions. Interestingly, the funds are paid back after
elections to farmers (with interest, so to speak) conditional on the respective
political mill chairman (or his party) winning office.
5. For instance, Gingerich (2014) takes advantage of a police investigation in Brazil into
an illicit campaign spending scheme which took place in the run-up to the elections in the
Brazilian state of Minas Gerais in 1998. Based on police reports, which contain detailed bank
transactions listing the names of those who received under-the-table election funds, he is able
to analyze the allocation of payments to local vote brokers and estimate their effect on election
outcomes. Mironov and Zhuravskaya (2014) aim to measure illicit payments by firms to politi-
cians in Russia. The authors find that firms involved in government procurement substantially
increase “tunneling” (defined as transfers by legitimate firms to fly-by-night firms established
with the purpose of taking cash out of companies) around regional elections. These illicit flows
exhibit a political business cycle, in contrast with firms not involved in public procurement.
Sukhtankar’s paper captures an important truth characterizing India’s

electoral dynamics: the costs of elections have grown so immensely in recent
years that politicians face incentives to recoup some of the financial “invest-
ments” made during the campaign by using their political positions to extract
rents. Indeed, political office is widely perceived to be a highly lucrative
proposition in India. Unfortunately, quantifying just how “lucrative” elected
office can be is not an easy task. However, some scholars have attempted
to estimate the financial returns to office, drawing on new data drawn from
affidavits candidates submit at the time of their nomination which detail,
among other parameters, their financial assets and liabilities. Analyzing the
affidavits of state and national incumbent legislators who won elections in
the early 2000s and then recontested elections several years later, Sastry
(2014) finds that the average wealth of sitting MPs and MLAs increased by
222 percent during their tenure in office (from an average of `1.8 crore in
the first election to `5.8 crore at the time of reelection).
More systematic explorations of the financial rewards to office that take
selection bias into account suggest more modest returns. Bhavnani (2012)
compares the change in winners’ and losers’ self-declared family assets in
the country’s two most recent state and national elections, using a regression
discontinuity design. His results indicate that the average election winner
increased his assets by 4–6 percent a year.6 Using a similar design, Fisman,
Schulz, and Vig (2014) focus on the subset of elections where both winner
and runner-up from the same constituency run in the subsequent election.
Their analysis reveals that incumbents enjoy a “winner’s premium” of
4.5 percent on average; the additional returns to ministers, and to incum-
bents who face-off against freshmen legislators, are even higher, 10 and
12 percent, respectively.
Thus, careful econometric analyses suggest the financial returns to elected
office are real, but perhaps not as abnormally large as one might expect.
Yet there are reasons to treat these estimates with caution: they only look at
reported income; the data are restricted to a small sub-sample of all candi-
dates; and one of the two studies (Fisman et al. 2014) focuses on politicians’
assets net of liabilities, a contentious decision since access to low-interest
loans in India is often conditioned by political connections (Cole 2009).
Finally, the new candidate affidavit data has also shed light on the direct
link between the issue of political finance and the criminalization of politics
in India. Of the 543 members of the 16th Lok Sabha elected in May 2014,
6. Bhavnani concludes that 4–9 percent of election winners appear “suspect,” since their
asset growth is greater than what they would have earned based on their salaries (and perks)
as lawmakers.
34 percent face pending criminal cases while 20 percent face charges of a

“serious” nature. The situation is broadly similar at the state level, where 31
percent of elected MLAs face pending cases (15 percent fall into the serious
category). To compound matters, the share of elected officials with pend-
ing criminal cases has been increasing, rather than decreasing, over time.
In 2004, 24 percent of MPs faced criminal cases (12 percent faced serious
charges). This proportion grew to 30 percent in 2009 (15 percent serious)
and 34 percent (21 percent serious) in 2014.
Research by Vaishnav (2012) has shown that one reason parties value
candidates with criminal records relates to their access to financial resources.
As the costs of elections have surged, parties, in response to their declining
organizational strength, have grown increasingly reliant on self-financing
candidates. In fact, there is a strong correlation between a parliamentary
candidate’s personal assets—a good proxy for financial capacity—and the
likelihood of winning election. The affidavit data convincingly show that
criminal candidates, in turn, have a distinct financial advantage over “clean”
candidates, controlling for a range of possible confounding factors.7
Dutta and Gupta (2014) have a slightly different explanation, but it too
emphasizes the importance of money. The authors present a formal model
which assumes that candidates facing criminal charges do face a certain
degree of negative stigma amongst the voting population. In other words,
voters will—all else equal—be less likely to vote for candidates under
criminal scrutiny. However, there are offsetting considerations. Since cam-
paigns are costly, candidates with wealth can draw upon their largesse to
win disaffected voters by convincing them of their “innocence.” In a related
vein, the authors argue that wealth offsets the electoral disadvantage criminal
candidates face on account of negative stigma. The authors validate their
hypotheses using data from the 2009 Lok Sabha election.
2.4. Public Sector Recruitment, Postings, and Transfers

A final proximate cause of corruption is the broken system of public sector
recruitment, postings, and transfers, which perpetuates—indeed, practically
mandates—corruption by public officials. There are at least two primary
reasons this system is dysfunctional. First, recruitment, transfers, and post-
ings are often conducted on the basis of bribes rather than merit. Second,
the wage and incentive structure does not adequately reward performance
and punish malfeasance.
7. A descriptive analysis by Sastry (2014) confirms this association as does a more sys-
tematic regression approach by Dutta and Gupta (2014).
It has long been known that recruitment, transfers, and promotions

across the administrative services at the state and central levels in India are
regularly made on the basis of bribes rather than merit (with the recruitment
of elite IAS officers serving as a prominent exception). Wade (1985) in
his aptly titled and seminal article “The Market for Public Office” blames
“the corruption-transfer mechanism and its effects on bureaucratic initia-
tives” for the failure of the Indian development state. At the very least, the
consequences of the allocation of public sector posts on the basis of money
rather than merit include a multiplier effect on corruption, since officials
who paid to obtain these posts must recoup their costs through rent extrac-
tion. For example, Wade documents that the cost of obtaining the post of
Superintendent Engineer in the Irrigation Department was 40 times the
average salary for the position. This leads to a vicious circle: the cost is
high because of the rents that can be extracted from the post, incentivizing
the employee to recover their investment once in office. A related further
consequence is that it is unlikely that honest officers could obtain these
lucrative posts.
Before a skeptic claims that the evidence we cite is 30 years old, let
us point out that the situation has not improved since. Krishnan and
Somanathan (forthcoming) note that, “the promotion of state service offic-
ers has been heavily politicized.” Using data on the universe of serving IAS
officers, Iyer and Mani (2012) empirically document the system of politi-
cal control over bureaucrats, again pointing out that transfers are the key
mechanism used. In one highly publicized recent case, the media reported
that IAS officer Ashok Khemka, who earned a reputation for fighting graft
(including lodging investigations against dodgy land deals involving Robert
Vadra, Sonia Gandhi’s son-in-law) has been transferred no fewer than 46
times in a 22-year career (Siwach 2015).
In addition to hiring, the system of remuneration of public sector employ-
ees leaves much to be desired. There are at least two problems. First, the
wage structure is compressed, with entry level wages too high relative to
private sector wages (see Figure 2 from Muralidharan and Sundaraman
2013) and top wages too low relative to comparable private sector compen-
sation. Second, there are virtually no rewards or punishments on the basis
of performance, and hence little incentive to perform well.
High entry level wages lead to too many people spending too much effort
trying to simply enter public service. This problem is exacerbated by rules
governing attempts at clearing the civil service entry examinations, with
current regulations allowing 4–7 attempts (depending on caste category)
between the ages of 21–30 (Krishnan and Somanathan, forthcoming). If a
candidate spends nine years outside the labor force simply trying to get a
public sector job, it is quite likely that this expenditure in terms of opportu-
nity cost and investment in studying must again be recouped. On the other
end of the spectrum, wages for top level administrators lag behind their
counterparts in the private sector, increasing the temptation to be corrupt.
For example, Krishnan and Somanathan point out that wage compression
(at least until the 1990s) “led to a loss of morale and an excuse and justifi-
cation for corruption.”
Finally, there is little reward to performing well, and worse, rarely any
punishment to nonperformance. The entrenched power of employee unions
and onerous government regulations means that, short of murder, it is nearly
impossible to fire a public sector employee. The news media recently
reported on the amazing case of a government official who did not show
up for his job for 24 years, yet continued to draw salary while numerous
attempts were made to fire him (Agence France-Presse, 2015). Krishnan
and Somanathan suggest euphemistically that “judicial interpretation and,
in particular judicial leniency to civil servants who perform inefficiently,
has reduced efficiency.” Under such circumstances, the only surprise is that
there remain honest public officials at all.
3. Varieties of Corruption: Magnitudes and Consequences
Having reviewed some of the leading drivers of corruption, in this section we

turn to a review of the literature on corruption in India, focusing on evidence
on the scale and scope of corruption and on the crucial issue of whether cor-
ruption actually affects economic and political outcomes. To make sense
of this vast literature, we follow a simple rubric. Since our definition of
corruption is based on actions of public officials, we categorize corruption
according to the nature of the corrupt actions taken by these officials.
Our first category involves facilitative corruption. This type of corruption
involves officials charging fees or bribes for activities that they should be
doing in the first place. For example, this category would include all types
of “speed money” bribes to obtain government services like ration cards,
passports, etc.
The second category is collusive corruption. This involves officials
breaking or bending rules to benefit bribers, and is particularly difficult to
detect since no party has an incentive to report the crime. Bribes paid to
bypass fines and regulations, kickbacks from procurement in government,
and bribes paid to illegitimately obtain government contracts or licenses

would all fit into this category.
The final category relates to extractive collusion. In this case, the official
simply extracts funds from the government or private parties, either through
harassment or stealth. Within this category we would include embezzlement
from public funds, harassment bribes, as well as actions like shirking or not
showing up to work.
Our categorization may not necessarily encompass every instance of
corruption by bureaucrats or politicians, although it does cover the majority
of cases in the literature. Furthermore, it provides us with a useful tool for
conceptualizing the impact of corruption. Hence, we not only summarize
the empirical literature on each category of corruption, but also attempt to
outline how each type of corruption might distort allocations.
A word of caution is in order prior to delving into the existing body of
corruption-related research: India is vast and heterogeneous, and many of
the studies conducted in one state or region may not be easily extrapolated
to other states or regions. When issues of external validity and small sample
size warrant particular attention, we highlight them. When multiple studies
in different contexts reach the same conclusion, we are more confident in
drawing policy prescriptions. In all other cases, we leave the reader with
this general caveat.
3.1. Facilitative Corruption

The first category is a form of corruption that a majority of Indians have
likely experienced: the payment of bribes to obtain routine government
services and documents such as ration cards, driver’s licenses, passports,
residence and caste certificates, etc. Transparency International notes
that 54 percent of urban respondents who had contact with nine common
government service organizations had to pay a bribe to obtain the service
(Transparency International 2011 South Asia Barometer). The popular web-
site www.ipaidabribe.com was started in part because of the commonality
of this type of experience, and although the self-reports collected there are
not representative, the site claims millions of visits and tens of thousands of
reports from 600 cities and towns across India. Popular resentment against
this type of corruption has led to the introduction of Right to Service legisla-
tion, although the bill is still languishing in the Lok Sabha.8
8. The bill’s full title is: the Right of Citizens for Time Bound Delivery of Goods and
Services and Grievance Redressal Bill, 2011.
Theoretically, this type of corruption likely comes under the heading of

“corruption without theft,” where officials pass on the official price of the
good or service to the government but charge additional fees or bribes that
they keep (Shleifer and Vishny 1993). Under this scenario, officials need to
artificially restrict the quantity of service provided so that they can charge a
higher overall price. Further, social and efficiency consequences are likely to
arise from the “wrong” people—from the point of view of society—getting
the document or service: e.g., unqualified drivers getting driving licenses,
rich people obtaining Below Poverty Line (BPL) cards, etc. Of course, if
bribes simply serve as user fees or “speed money” to incentivize bureaucrats
to work faster but do not allow “bad” types to obtain services or documents,
the negative consequences might be mitigated.
Recent empirical work provides us with good evidence on the extent and
consequences of this type of corruption. Bertrand, Hanna, Djankov, and
Mullainathan (2007) followed 822 applicants for driver’s licenses in Delhi.
As predicted by theory, bureaucrats artificially restrict licenses and create
red tape in order to charge applicants more than official fees and clear the
market: government officials seem to arbitrarily fail applicants taking the
official driving test, as the authors find that failure on this test is uncor-
related with actual driving ability as measured by an independent driving
test. Accordingly, applicants must make multiple trips to obtain licenses,
and end up paying 2.5 times the official fees for the license.
These bribes do not simply represent a transfer from applicants to bureau-
crats, but are actually harmful to society: 71 percent of license getters do
not take the licensing exam, and most damningly 62 percent of license get-
ters failed the independent driving test. Further, the authors experimentally
manipulate willingness to pay (private value) by offering a random subset
of study subjects a substantial bonus if they obtain their licenses quickly.
They find that the licensing process is very responsive to private value, but
not to social value: those offered a bonus were much more likely to both
get a license but also to be unable to drive when compared to the control
group. At this point, bribes to obtain licenses cross over from being simply
facilitative corruption to collusive corruption (discussed below).
A similar story of extra-statutory fees and the “wrong” people getting
government services and benefits holds true for BPL cards. These cards
entitle households to a range of welfare benefits, most importantly to subsi-
dized food under the Targeted Public Distribution System (TPDS). Niehaus
et al. (2013) surveyed 14,074 households in rural Karnataka to learn about
the process for obtaining BPL cards. They found that bribery is widespread:
75 percent of households reported paying bribes to obtain the cards, although

the average payment above official fees was small: `14.
More importantly, however, they found that 48 percent of households
are misclassified. Seventy percent of households that were ineligible to
receive BPL cards—based on criteria such as owning a vehicle, color TV,
gas connection, or more than 5 acres of land—had a card (type 1 error), and
worryingly, 13 percent of eligible households did not (type 2 error). Overall,
statutory eligibility was much more strongly correlated with income than
actual ownership of cards, suggesting that reasonable targeting rules were
perverted by the corrupt allocation process.9
Two studies in Delhi (Peisakhin and Pinto 2010; Peisakhin 2012) found
similar arbitrariness and restriction in the provision of ration cards (required
for TPDS benefits) as well as voter ID cards, although these results must be
viewed with caution given the small and nonrepresentative samples (86 and
121 individuals in one slum and university area, respectively). For study
subjects, it was practically impossible to obtain these documents without
paying a bribe or resorting to a Right to Information request. We discuss
these studies further in Section 4.
3.2. Collusive Corruption

Collusive corruption occurs when officials, conspiring with bribers, bend
or break rules and regulations. Assuming that the given rule was good for
society in the first place, distortionary effects can be large. For example,
government agents might grant firms contracts in exchange for bribes or
kickbacks, which could result not just in economic distortions but also
efficiency losses. If an agent awards an infrastructure contract to a firm for
reasons independent of firm quality, it is possible that the firm is unqualified
to execute the contract faithfully or will shirk in order to recoup the financial
loss incurred by the bribe payment. A 2014 survey of “Global Economic
Crime” by the accounting firm Pricewaterhouse Coopers LLP (PwC) found
that the industry reporting the greatest degree of procurement fraud was
government/state-owned enterprises (PwC 2014). In India, corruption in
public procurement is a well-identified obstacle to improving the country’s
investment climate (UNODC 2012).
9. Besley et al. (2007) examine data from four south Indian states and find that while
local-level politicians target BPL cards to households which are relatively disadvantaged
on average, households in which politicians themselves are living are much more likely to
possess BPL cards.
Unfortunately, this type of corruption is also difficult to empirically

document, since neither the bribe-giver nor bribe-taker has an incentive to
publicize this transaction (Bardhan 1997). However, a recent paper by Duflo
et al. (2013) breaks new ground. In most markets in which the state plays
a regulatory function, regulated firms themselves often choose and pay for
the “third-party” audits meant to monitor compliance. This naturally creates
a conflict of interest: the firm has an interest in an audit that paints the firm
in a good light, while the auditor has an interest in satisfying the client in
order to maintain business. This creates an incentive for rampant corrupt
behavior although it need not necessarily have an adverse direct impact on
economic efficiency.
In the context of a third-party audit experiment involving pollutant-emit-
ting plants in the state of Gujarat, the authors uncover systematic evidence
of corruption in the audit reporting for plants in the control group—that is,
those plants that are audited by firms selected and paid for by the plants
themselves. When auditors were hired and paid by the firms they were
auditing, 29 percent of auditors falsely reported pollution below the regu-
latory standard (even though actual emissions were above the standard).
For control group emitters, auditors reported that only 7 percent of plants
violated the government standard, when in reality 59 percent were emitting
more than the standard.
To our knowledge, only one rigorous empirical study documents cor-
ruption in public sector procurement practices. In the context of a study on
the introduction of electronic procurement, Lewis-Faupel, Yusuf, Olken,
and Pande (2013) examine the tendering process for manual procurement
drawing on a random sample of road contracts issued by the state of Uttar
Pradesh. As the authors note, manual procurement is subject to corruption
on several grounds. Because documentation is not public and exists only in
the written form, the government can provide private information to favored
bidders that would give them a competitive advantage or use its discretion-
ary authority to disqualify bidders on spurious grounds. Indeed, the authors
find there is little competition for public sector road-building contracts
because many firms are disqualified on “technical” grounds. In 95 percent
of cases, the government only ended up evaluating a single firm’s financial
bid. Where there are multiple initial bidders, in the case of any technical
disqualification, the authors report that “all but one bidder are disqualified
100 percent of the time.” The pattern of disqualification is consistent, the
authors argue, with corrupt officials rigging the procurement process to
favor a pre-determined winner.
At least one study (Kapur and Vaishnav 2015) has drawn a link between
collusive corruption and India’s broken system of financing elections.
The motivating premise of their study is that politicians often turn to

private firms for illicit election finance in sectors where the discretion-
ary powers of the state are large, in line with our discussion of one of
the two “deep” drivers of corruption in India. Firms operating in highly
regulated sectors are natural targets as election donors since politicians
can exchange policy discretion or regulatory forbearance for campaign
contributions.
The authors specifically focus on the role of the construction sector,
which depends heavily on the availability of land, an input that is tightly
controlled by state authorities. Kapur and Vaishnav hypothesize that builders
operating in the sector will experience a short-term liquidity crunch as elec-
tions approach because of their need to re-route liquid funds to campaigns
in the form of election payments. Using a novel monthly level dataset that
captures variation in the state-wise demand for cement—the indispensable
ingredient of the modern construction sector—the authors confirm the
presence of an electoral cycle in cement consumption in India, confirming
the original prediction. Consistent with their theoretical predictions, the
negative shock in cement consumption is more intense for state elections
(states have primary regulatory responsibility for land), urban states, and in
especially competitive elections.
Finally, there is a burgeoning literature on politicians manipulating the
targeting of goods, services, licenses, jobs, or other transfers in order to reap
electoral benefits (see Golden and Min 2013 for a broad review). Whether
these activities are “corrupt” in a narrow sense can be debated; distribu-
tive transfers in this realm are variously referred to as pork, clientelism,
or patronage. While such transfers can entail a misallocation of public
resources, we consider these as part of the political process, and difficult to
define as “corruption,” thus do not explore the literature in this area.
3.3. Extractive Collusion

This third category of corruption is variously described as leakage, diver-
sion, or embezzlement. In its simplest form, the government tries to send
benefits of some kind (money, food, medicine) to recipients, and officials
in charge of delivery simply steal them rather than delivering them to the
poor. Late Prime Minister Rajiv Gandhi once famously estimated that only
15 percent of benefits disbursed by the government of India actually reach
the poor. In addition, overbilling the government for benefits in the name of
fake recipients also falls under this category. Thus, theft can be from both
beneficiaries, which directly harms them (underpayment), and from the
government, which harms taxpayers in general (overreporting).
Embezzlement has distortionary consequences for optimal public finance.

First, it might make seemingly progressive public programs regressive, if
officials are generally richer than beneficiaries and taxpayers (Olken 2006).
Second, it affects optimal rules for the allocation of public funds: without
corruption, governments would simply equate the marginal social costs of
raising funds to the marginal social benefits of delivery, but with embez-
zlement the marginal social benefits need to be adjusted by the fraction of
funds that actually reach the poor (Niehaus and Sukhtankar 2013b).
In addition, leakage may have systemic negative consequences too:
officials may allocate time towards activities focused on embezzlement
(Murphy et al. 1991) rather than implementing public programs as they are
meant too. However, one line of argument suggests these rents may keep
officials incentivized to implement programs.
Recent empirical work provides evidence on leakage from at least the
two largest welfare programs, NREGS and TPDS. The methods used in
these studies are straightforward: they involve comparing official records of
disbursements of benefits against beneficiary surveys. Of course, beneficiary
recall and misreporting are concerns with this methodology, so precise levels
must be viewed with caution.
Niehaus and Sukhtankar (2013a, 2013b) surveyed about 3,000 listed
NREGS beneficiary households in three districts in Odisha and one in
Andhra Pradesh (AP), comparing official records of disbursements of
NREGS wages against beneficiary reports in original surveys. The results
are disheartening: about 70–80 percent of the NREGS labor budget is embez-
zled before it gets to beneficiaries. This corruption directly hurts beneficiar-
ies, as the work they do is not correctly remunerated, and their wages are
underpaid. It also hurts taxpayers, through over-reporting of work done, as
the exchequer pays out far more than intended.
In addition, government efforts to increase benefits—the statutory wage
in this case—are entirely thwarted, as none of the increase is passed on to
beneficiaries. The authors show that with this kind of corruption, a program
that is meant to set market wages instead ends up being a price-taker: ben-
eficiaries are just paid the prevailing market wage in the area.
It is important to keep in mind that these results, although representative
for the areas surveyed, correspond to districts that are likely more backward
and corrupt than the median district in India.
Khera (2011) highlights heterogeneity across India in embezzlement from
public programs while examining diversion of food grains from the TPDS.
Comparing state-level offtake for TPDS (i.e., the amount of grains that
states obtain from the Food Corporation of India (FCI)) of rice and wheat
to NSS survey reports of grains received by beneficiaries from Fair Price
Shops (FPS, or ration shops), she finds that the overall rate of diversion in
India in 2007–08 was about 44 percent. Estimates range from essentially
no diversion in Chhattisgarh to almost 90 percent diversion in Bihar. These
estimates are likely to be the upper bound of pure leakage, since some of
the grains that do not reach beneficiaries may simply be due to losses in
transport or spoilage or other mismanagement.
These estimates are very close to the Government of India’s own esti-
mates of diversion in the TPDS. A report by the Planning Commission
(Programme Evaluation Organization 2005) finds that 58 percent of food
grains issued by the FCI do not reach the poor (defined as BPL families),
which is comparable to the 54 percent figure that Khera estimates for
2004–05. The report also monetizes the magnitude of the loss and the cost
of delivery, calculating that for every Rupee transferred to the poor, the
government spent `3.65. In other words, the poor obtain only 27 percent of
the benefits they are meant to receive.
In addition to pure theft, an insidious type of corruption involves public
sector employees not showing up to work when they are supposed to, or,
more broadly, shirking on the job. It is, in a way, similar to embezzlement,
since it effectively equates to theft of time from the government. However,
the causes, consequences, and strategies to combat this form of corruption
differ significantly. Absence or shirking on the job is common in the public
sector across the world, since disciplining public sector employees proves to
be difficult given strong unions and other political economy factors.
The consequences of absence could include long-run harm to human
capital, mainly through education and health, in the country. As Chaudhury,
Hammer et al. (2006) point out, absent health and education workers basi-
cally mean closed hospitals and schools in developing countries, since there
are no substitutes and many of them have single providers. The unpredict-
ability of absence may also discourage users from attempting to access
these services in the first place. Moreover, alternatives to public schools
and hospitals may be too expensive and/or just as ineffective.
Chaudhury et al. (2006) conducted a representative survey across India to
measure absence amongst teachers and health care workers in government
schools and health centers. After conducting random checks of schools and
clinics during working hours, they found that the rate of absence of teachers
is 25 percent and health care workers 40 percent. This compares to 5 percent
in developed countries. Other work corroborates these absence findings in
smaller samples: Banerjee et al. (2010a) find 27 percent teacher absence
in Jaunpur district in Uttar Pradesh, and Banerjee, Duflo, and Glennerster

(2008) find 54 percent health care worker absence in Udaipur in Rajasthan.
An updated survey conducted by Muralidharan et al. (2014b) returned to
the same schools surveyed by Chaudhury et al. (2006), and found that little
had changed: teacher absence rates were 23 percent across India.
What is worse is that these absence rates are likely a lower bound, since
Chaudhury et al. were conservative in what was counted as an absence:
e.g., they did not count part-time employees, and took the head of the school/
clinic at her word if she said an employee was not supposed to be work-
ing. In addition, the authors also find that even when teachers are present,
they are teaching only 45 percent of the time. This is in spite of also being
conservative in defining teaching activity to include any time a teacher was
in the classroom. Similarly, health care worker absence is compounded by
the fact that even when they are present, public doctors treat patients much
worse in public clinics than they do in their own private practices (Das and
Hammer 2007).
The depressing news continues, as Chaudhury et al. also find that absence
rates are higher in poorer areas: a doubling of per capita income is correlated
with absence rates that are 6 percentage points lower. Health care also suf-
fers: the quality of care is such that in poor areas, unqualified private doc-
tors tend to perform better than qualified public doctors (Das and Hammer
2007). Further, hiring additional teachers only leads to greater absence:
hence effective marginal rates of absence were even greater than average
rates (Muralidharan et al. 2014b).
The causes of education and health worker absence are fairly straightfor-
ward to understand. Given political economy constraints, these civil servants
rarely, if ever, face any sort of punishment for shirking. Chaudhury et al.
found only one instance of an employee firing out of 3,000 cases where a
teacher was absent. Further, they also find that in more powerful positions,
occupants were more likely to shirk: e.g., doctors were more likely to be
absent than nurses, men more likely than women, and head teachers more
likely than ordinary teachers.
The consequences of absence are, of course, poor health and education
outcomes. For example, while enrollment rates are almost 100 percent
amongst children aged 6–14, only 56 percent of children in rural India
can read a simple story by grade 5 (Chaudhury et al. 2006). Duflo, Hanna,
and Ryan (2012) find that reducing teacher absence from 42 percent to 21
percent in rural Rajasthan improved student test scores by 0.17 standard
deviations, which is a very large effect in this literature. Finally, there is a
large fiscal cost to the government: Muralidharan et al. (2014b) estimate

this to be $1.5 billion a year, or 60 percent of the revenues raised by the
special education tax.
4. Strategies and Legislation to Combat Corruption:

What Might Work?
The previous section catalogued the categories of corruption scholars have

rigorously studied in India and their impacts on society more broadly. In this
section, we turn our attention to solutions. We begin with a categorization
of broad strategies to combat corruption. We use this schema to classify the
current set of reforms, and to examine what evidence there is on the potential
effectiveness of each strategy.
The broad categories of tools to combat corruption are as follows:
1. Information/ bottom up monitoring

2. Technology
3. Financial incentives/ performance pay / efficiency wages
4. Electoral reform/ political incentives
5. Legal reform
6. Policy reform
4.1. Information/Bottom-up Monitoring

Information is seen as a basic pillar of the fight against corruption worldwide.
According to Transparency International, “access to information and a strong
civil society are essential for good governance and public accountability.”10
Remarkably, India’s RTI Act (RTIA) that was passed in 2005 is ranked as
the second best right to information law in the entire world by the Center
for Law and Democracy, placing it above every single OECD nation in
this regard.11
Of course, having a law on the books is one thing, and implementing it
is quite another. India also has laws that guarantee employment for rural
10. http://www.transparency.org/whatwedo/activity/action, accessed June 25, 2014.

11. The universe of eligible laws is not limited to developing countries, or emerging
nations, but only to countries that have an existing law on the books. The ranking is based on
overall score in the categories of Right of Access, Scope, Requesting Procedures, Exceptions
and Refusals, Appeals, Sanctions and Protections, and Promotional Measures: http://www.
law-democracy.org/live/wp-content/uploads/2012/08/Chart.pdf, accessed June 25, 2014.
dwellers (NREGA), access to basic education (Right to Education), and the

right to food (National Food Security Act), yet their implementation leaves
much to be desired. The good news, however, is that two very similar stud-
ies suggest encouraging effects of the RTIA on citizens’ ability to obtain
public services.
Peisakhin and Pinto (2010) examined the process of obtaining ration
cards in one slum in Delhi. The study randomized 86 applicants into four
groups: the main treatment group was assisted in preparing RTIA requests,
and outcomes of this group were compared to groups who either: (a) paid
a middleman to obtain cards through (ostensibly) bribing officials; (b) pre-
sented a letter of support from an NGO; and (c) a control group that simply
submitted the necessary paperwork. The RTIA request in the intervention
asked the relevant official “for information about the status of their applica-
tion and about the average time for ration card applications in this district.”
The letter from the NGO noted that “the application was eligible for a ration
card and urged prompt administrative action.”
Remarkably, 94 percent of applicants in both the bribe and RTIA groups
received their ration cards within a year, while only 21 percent received them
in the NGO and control groups. The bribe and RTIA groups also both had
much shorter median processing times: 82 days for the bribe, 120 days for
the RTIA, and 343 days for the control.
A related experiment compared RTIA requests, bribes, and a control
group in the process of registering 121 applicants for a voter ID card, with
an additional element that distinguished between applicants from a middle-
class university area and in a slum in Delhi. Again, the bribe and RTIA
groups performed significantly better than the control group, in both poor
and middle class areas.
The very small sample sizes and the fact that the studies were done in
small urban areas in Delhi means the results must be viewed with caution.12
While the results are encouraging, the mechanism through which the RTIA
works for these requests is not clear. In particular, it hardly seems like the
binding constraint for receiving the service is “information on the status
of the application,” and further, penalties for noncompliance with RTIA
requests are extremely rare and not very large. Interestingly, the authors
themselves speculate that the RTIA requests work “not so much because
of the Act’s penalty provisions, which are rarely used, but rather because
in India’s ultra-competitive bureaucracy, any blemish on a public servant’s
career can negatively affect his chance of promotion.”
12. For example, Roberts (2010) reports that awareness of the RTIA in urban areas is three
times that in rural areas.
Thus, it appears that the RTIA is less about information than shifting the
bargaining power to applicants, and giving them a threat which they can
use credibly to potentially sanction the official. This interpretation fits our
framework above, and also helps explain why no other study which provides
information to applicants or beneficiaries finds any effect of providing this
information, whether in public works or education.
In the case of NREGA, e.g., Ravallion et al. (2013) conducted a large-
scale randomized trial in Bihar to test for the effects of providing infor-
mation to actual and potential program participants on their rights under
the act. Baseline levels of information about and participation in NREGS
were very low. While 56 percent of the rural population was BPL, only
about 10 percent participated, even though more than 75 percent of those
surveyed—5,000 rural individuals in 150 villages spread across rural
Bihar—had heard about NREGS. In the control group, only 22 percent knew
how many days they are allowed to work, 32 percent knew the wage rate,
29 percent knew about receiving unemployment insurance if work is not
provided within 15 days of applying, and almost no one knew that applicants
are supposed to get work within 15 days.
To test whether this situation could be remedied, the authors showed a
randomly selected treatment group of villages an educational, emotionally
captivating film about NREGS to the village. The intervention increased
knowledge of the program: a 53 percent increase in knowing about number
of days allowed to work, a 36 percent increase in people who know about
wage rate, a 33 percent increase in knowing about unemployment insurance,
and a 70 percent increase in knowledge of getting work within two weeks.
Despite this increase in awareness, however, reported survey outcomes on
actual or desired participation, wage rates or days worked remained entirely
unchanged. While the NREGA guarantees employment, on average partici-
pants said they would like 44 more days of work, so rationing was evident.
These results are also consistent with evidence from Odisha presented
in Niehaus and Sukhtankar (2013b). They survey a large sample of listed
NREGS participants in three districts in Odisha, and ask about days worked
and wages received around a time during which the official minimum wage
rate increased from `55 per day to `70 per day. While over 80 percent
of respondents knew about the wage change, and over 70 percent could
accurately name the new wage, on average almost no one working on
NREGS received a wage increase. The only exception to the rule was that
villages where an NGO was active received modestly higher wages after
the wage change, perhaps because these NGOs helped keep local officials
accountable.
Providing information as well as encouraging community-based monitor-

ing was also largely ineffective in other contexts in India. Community-based
monitoring is fast becoming a buzzword in the development community
around the implementation of public programs (Björkman and Svensson
2009). The enthusiasm derives from the idea that local communities have
easier access to information regarding the performance of local government
officials, and hence involving these communities in program management
might improve the functioning of programs. Worldwide, there are mixed
results on the performance of community-based monitoring, with Björkman
and Svensson (2009) finding improved health outcomes in Uganda but Olken
(2009) finding no effects on public works in Indonesia.
In the context of NREGS, “social audits” have been touted as a means
to tackle corruption. The state of AP is the leader in implementing these
audits, in which a team of district and state auditors train local villagers in
auditing techniques and conduct a week-long exercise together whereby
official expenditure records on NREGS are tracked in villages. At the end
of the exercise in each subdistrict, a public hearing is held in the presence
of NREGS officials as well as local beneficiaries, and complaints and tes-
timonies are read out.
Afridi and Iversen (2013) analyzed social audits in AP by collecting a
panel dataset on three rounds of social audits in 300 villages across 8 districts
between 2006 and 2010. They find that these audits largely do not have
a deterrent impact on corruption: having an initial audit does not change
the number of easy-to-detect irregularities in round two, and it actually
increases the number of hard-to-detect irregularities. There are no effects
on other important program outcomes such as employment generation. The
authors suggest that a leading explanation for the lack of effect is that once
discovered, official malpractice is rarely punished.13
In rural Uttar Pradesh, village education committees (VECs), which
include parents of children in local schools, are mandated by law to serve
as intermediaries between the village and educational authorities. Banerjee
et al. (2010a) conducted a randomized trial in 280 villages in the Jaunpur
district, in which an NGO provided information on the roles and responsi-
bilities of the VECs, invited villagers to create “report cards” on the status
13. Less than one percent of corrupt officials are actually removed from office or face
serious action; even mild punishments such as suspensions are meted out in fewer than three
percent of cases; and amazingly officials seem to be able to get away scot-free with their
earnings, since over 87 percent of embezzled amounts were not recovered three to seven
years after the audits.
of education in villages, and trained volunteers in teaching basic reading

skills to primary school children. None of these treatments were successful
in encouraging increased involvement in public schools.
Overall, our assessment is that information and encouragement of com-
munity participation, on their own, will have little effect in addressing cor-
ruption. Instead, putting more bargaining power in the hands of beneficiaries
and applicants is more likely to result in greater impact. In this regard, the
various “Right to Service” bills which have been passed in many states and
are being considered in several others as well as Parliament, have the poten-
tial to reduce corruption in public service delivery provided they include
effective penalties for misbehaving officials. As of January 2012, at least
nine states had passed “Right to Service” laws. While the specific provisions
of the respective states laws differ, they all guarantee the right to a specified
list of services to citizens in a time-bound fashion and institute penalties
for government officers who fail to comply (see Raha 2012 for a review).
A central act, The Right of Citizens for Time Bound Delivery of Goods
and Services and Grievance Redressal Bill, was introduced in 2011 and is
pending in the Lok Sabha (PRS Legislative Research 2012). This bill in, its
current form, seems promising. It forces public authorities to detail timelines
for delivery for all goods and services that it provides, and allows citizens
to file complaints if these are not delivered on time, or if they experience
misconduct from a government officer of the authority. The bill also man-
dates the appointment of a Grievance Redressal Officer (GRO) and State
and Central Grievance Redressal Commissions, which can judge appeals.
Complaints must be addressed within 30 days, and multiple appeals are pos-
sible to higher authorities. Most importantly, the bill allows for fines up to
`50,000 on the GRO or officers guilty of misconduct, and also allows some
of this fine to be awarded as compensation to the complainant.
These provisions of the bill have the potential to both reduce ability of
bureaucrats to arbitrarily restrict public services in order to extract bribes,
and increase bargaining power of applicants for public services. Of course, it
remains to be seen how central and state government authorities and officials
respond to the provisions, whether loopholes can be exploited, and whether
fines and penalties are actually assessed.
The recent history of the RTIA provides us with some cautionary fodder.
A 2010 review of the law suggested that the “use of the law has been con-
strained by uneven public awareness, poor planning by public authorities,
and bureaucratic indifference or hostility” (Roberts 2010). Furthermore,
recent attempts have been made to reduce the number of entities under its
purview, most importantly the Central Bureau for Investigation (CBI) and
political parties.14
While these attempts prove that the act has some bite, actual cases of
penalties and sanctions imposed are rare. For instance, several studies
(quoted in Roberts 2010) have found that State Information Commissions are
often reluctant to levy penalties on noncomplying officers since that would
unfairly penalize officers who lack training or experience or who have to
deal with systemic shortcomings. Meanwhile, anecdotal evidence grows
that the RTIA can be misused for competitive or personal gain; overall, a
rigorous analysis of the impact of this act is imperative.
A final transparency mechanism, which does not yet exist but is awaiting
parliamentary approval, is a central procurement portal, embedded within
the Public Procurement Bill (2012).15 The bill mandates the establishment
of a “Central Public Procurement Portal,” which would serve as a reposi-
tory for materials related to government procurement, such as documents
related to prequalification, registration and bidding, as well as participation
details and final decisions or appeals. While such a move is a good first step,
a real breakthrough could be achieved if the government were to consider
publishing actual government procurement contracts. Kenny and Karver
(2012) have argued that such “Publish What You Buy” provisions can not
only reduce corruption but also lower the costs of contracting to the benefit
of governments, contractors, and citizens.
4.2. Technology
Given the emergence of information technology services as a dynamic sector
of the Indian economy, state and central governments have often looked
towards technology as a potential silver bullet for tackling corruption. The
recently defeated UPA government embarked on an ambitious initiative—
Aadhaar—to deliver biometrically authenticated unique IDs to all residents
of India. Authorities believed that this initiative would revolutionize the
delivery of government services, with the Unique ID Authority claiming
that “Aadhaar will empower poor and underprivileged residents in accessing
services such as the formal banking system and give them the opportunity
to easily avail various other services provided by the Government and the
14. Parliament exempted the CBI from the RTI in 2011, while political parties are in a
pitched battle with the CIC, which ruled parties to be “public authorities” subject to RTI in
June 2013. We discuss the latter dispute in greater detail in Section 4.4.
15. The bill is pending before the Lok Sabha. More detail can be found in PRS Legislative
Research (2013).
private sector.”16 Former UPA Finance Minister P. Chidambaram called it

“a game changer for governance” (Harris 2013). After some murmurings to
the contrary, the NDA government, shortly after taking power in May 2014
decided to fully implement Aadhaar.
State governments have also embarked on their own initiatives to deliver
services, in particular online services for applying for ration cards, certifi-
cates of various kinds, obtaining land records, etc. (commonly referred to
as “e-sewa” initiatives).
The academic evidence on the impact of technology to reduce corruption,
however, is somewhat mixed across various sectors. For example, on the
e-sewa initiatives, Bussell (2012) writes that “while there was considerable
initial enthusiasm to use new technologies, the actual benefits offered to
citizens are constrained in many cases by persistent efforts to retain access
to a rich source of corruption: the bribes citizens pay to get the services they
are promised by the state.” Nonetheless, the studies provide some obvious
and clear lessons that can be used to inform future policy.
A first set of interventions using technology involve attempts to reduce
shirking and absence by teachers and health workers. Duflo et al. (2012)
worked with an NGO in the Udaipur district to employ digital cameras to
monitor attendance, along with financial incentives to reduce absence, in
single-teacher rural schools run by the NGO. Baseline absence rates of
44 percent were higher than all-India averages. In the randomly assigned
treatment group, teachers were given tamper-proof cameras and asked to
have students take date-stamped photos of the teacher and other students at
the beginning and end of each school day. This provided proof of attendance,
and teacher salaries were based on the number of days attended.
Comparing attendance for teachers who were given cameras and incen-
tives with a control group who did not receive either, the authors found
that absence was reduced by 50 percent in the treatment classrooms. More
importantly, teacher attendance actually translated into improved educational
outcomes for students: test scores were 0.17 standard deviations higher in
the treatment group after a year of the program.
Unfortunately, the promise shown by this intervention has not been rep-
licated in other settings. Banerjee et al. (2008) attempted a similar interven-
tion with nurses in government-run rural health centers, also in Rajasthan.
Again, baseline absence rates were very high, at around 60 percent. In this
case, assistant nurse midwives in the treatment group were given locked
16. http://uidai.gov.in/index.php?option=comcontent&view=article&id=58&Itemid=106,
accessed 3 October 2013.
and password-protected time-stamp machines and asked to stamp cards at

various points during the day, again with financial incentives for attendance.
The expectations for attendance were very low: the only monitored days
were Mondays, and pay was docked only once attendance dipped below
50 percent.
Initially, the intervention appeared to be successful, as absence rates
almost halved. However, nurses soon found a way to co-opt the system.
Official excuses for absence started going up, and 16 months after the
program started, there were no differences between treatment and control
group attendance.
A similar program attempted by Dhaliwal and Hanna (2014) in Karnataka
also showed the limitations of technology. While initially outcomes
improved, eventually the state health system found it increasingly difficult to
attract nurses to work at rural outposts. This suggests that another constraint
is the lack of human capital in rural areas.
One of the main differences between the teacher and health worker studies
was that the teachers worked for NGO-run schools, and the NGO was will-
ing to enforce penalties for absence. Governments face political economy
constraints arising from the power of teacher and health care worker unions
and their ability to help local politicians win elections, and hence interven-
tions that succeed in small scale pilots may not easily scale up when run by
the government (Acemoglu 2010).
For instance, Bold, Kimenyi, Mwabu, Ng’ang’a, and Sandefur (2013),
e.g., conducted a contract teacher intervention in Kenya where half of the
sample was devoted to a trial run by an NGO and the other half run by the
government, and find that while outcomes improved under the NGO, they
did not under the government. In sum, larger systemic constraints matter.
Other technological incentives, even those run by the government,
show more promise. Muralidharan et al. (2014a) study the Andhra Pradesh
Smartcard initiative, which used biometrically authenticated Smartcards to
make payments under NREGS and Social Security Pensions (SSP), the latter
of which makes monthly payments to elderly, widowed, and disabled indi-
viduals. Smartcards were a functional precursor to the integration of UID/
Aadhaar with these programs. The evaluation was conducted in partnership
with the Government of AP using one of the largest randomized controlled
trials ever done, and featured a randomized rollout of the program across
19 million beneficiaries, which enabled an empirically rigorous evaluation
of the new payments system.
Previously, payments were made in cash to beneficiaries, often by the
same officials who implemented these programs. The new system made
payments through local customer service providers (CSPs) who were

employees of banks contracted by the government to manage payments,
and added fingerprint authentication to ensure that actually intended bene-
ficiaries received the money. This large-scale initiative faced both vast
implementation challenges and pushback from local officials who stood to
lose rents. Despite the best efforts of the government, only about 50 percent
of payments in treatment areas were made via Smartcards after two years.
Nonetheless, the results are extremely promising. The poor gained sig-
nificantly from the reform, through improvements in the payment processes
that reduced both the time to collect payments and delays in transferring pay-
ments. There was also a significant reduction in NREGS and SSP leakage,
about 40 percent for both programs, with most of the effect coming from
reductions in overcharging the government for benefits that recipients never
received. Despite the reduction in leakage, there was no reduction in access
on either program. Smartcards were highly cost effective: time savings to
beneficiaries alone exceeded the entire cost of the program for NREGS;
further, the reduction in NREGS leakage was nine times greater than the
cost of program implementation. Finally, beneficiaries strongly supported
the program: 87 percent of NREGS and 92 percent of SSP beneficiaries
prefer the new payment system over the old one.
These results are highly relevant to understanding the likely impacts
of UID-integrated benefit transfers in India and similar programs in the
developing world, and suggest that investing in secure authentication and
payment infrastructure can significantly enhance “state capacity” to effec-
tively implement a broad range of programs.
Given our framework, the reasons why this intervention worked become
clearer. First, the electronic benefit transfer through CSP bypasses the local
officials who are the source of hold-up and asymmetric information, or at
least forces them to collude with CSPs. Further, live biometric authentica-
tion forces the presence of beneficiaries while collecting money, increasing
their bargaining power. Previously, officials could simply collect money on
behalf of beneficiaries even when they were not around (or in some cases
did not even exist).
Finally, using technology in government procurement also shows some
promise. A study by Lewis-Faupel et al. (2013) examines the impact of
electronic procurement (henceforth, e-procurement) on public works projects
in India and Indonesia. As the authors point out e-procurement seeks to
address three common shortcomings associated with standard procurement
practices: information asymmetries, collusion among bidders, and corrup-
tion. There is, however, the possibility of negative impacts should there be
large variation in access to Internet technology or the continued existence

of coercive tactics by powerful firms.
In India, the authors examine procurement practices between 2000
and 2006 in the central government rural roads scheme, Pradhan Mantri
Gram Sadak Yojana. Although the authors only have observational data,
they rely on the phasing in of e-procurement over time, which allows for a
difference-in-difference empirical strategy. In both countries, the study finds
that e-procurement actually increases the probability of the winning bidder
coming from outside of the region where the project is to be implemented.
This suggests that e-procurement improves competition in the market given
that it reduces the barriers to entry for firms to participate without being
physically present.
However, the authors found no statistically significant evidence that
e-procurement lowered prices paid by the government in India. They did
find declines in Indonesia, although they were modest in size and not statisti-
cally significant. When it comes to quality improvements, e-procurement
yielded positive gains on at least one measure in India. On the one hand,
the authors found no evidence in India of an improvement or reduction in
project delays although they found large declines in Indonesia. However,
for India they also had access to data on an independent audit on construc-
tion quality. E-procurement was responsible for a 13 percent improvement
in quality grades compared to projects with standard procurement norms.
In sum, e-procurement did not drive prices down through greater eco-
nomic competition. It did, however, improve competition by attracting
new firms—often from other regions—which were often of a higher qual-
ity. Furthermore, there was a positive impact on the actual quality of road
construction, as assessed by an independent quality audit.
Overall, the lessons from technological interventions are similar to
those from the information interventions, with one important caveat.
Technological interventions that rely on enforcement from higher authori-
ties will necessarily be constrained by political economy considerations,
but those that simply bypass middlemen bureaucrats or those that transfer
bargaining power to beneficiaries will be more likely to succeed.
4.3. Financial Incentives/Performance Pay/Efficiency Wages

While not quite as fashionable as technological interventions, financial
incentives are one of the most straightforward methods that governments
might use to combat corruption. Incentives could include simple bonuses
based on outcomes or penalties if public officials are caught engaging in
corruption. Indeed, many of the technological interventions described in the
previous section are combined with financial incentives, and, thus, we do not
discuss these studies in this section. The main lesson from these studies is
straightforward: high-powered incentives work, as long as they are enforced.
While financial or performance-based incentives for undertaking specific
actions may indeed work, there is also a potential concern that they may also
lead to negative consequences. For example, nurses may now show up to
work but put in less effort at work. Teachers paid on the basis of student test
scores may “teach to the test” at the expense of encouraging “real” learning.
Fortunately, evidence from a large-scale randomized experiment of gov-
ernment teachers in AP suggests that these types of concerns are perhaps
overblown (Muralidharan and Sundaraman 2011). Teachers in 300 govern-
ment schools in AP were randomly allocated incentive programs that gave
them bonuses of up to 3 percent of their salary based on the performance of
students on tests. After two years, student test scores improved significantly
in the subjects whose test results counted for bonuses: math and language.
Importantly, however, test scores also improved in subjects whose test
results did not count towards bonuses—science and social studies —suggest-
ing that enhanced teaching effort spilled over into these subjects. Moreover,
results were positive across the distribution of students, suggesting that
teachers did not just concentrate on the best or worst students.
In addition to the specific incentives described above, there is a long-
standing idea that government officials need to be paid more overall: they
are only corrupt because they are so poorly paid and need to supplement
their income through illicit activities. More generally, the idea that “effi-
ciency wages” could prevent corruption is very old, going back to at least
the seminal work by Becker and Stigler (1974). In practice, however, there
is little evidence that such measures will be effective.
Niehaus and Sukhtankar (2013a) indirectly test if efficiency wages mat-
ter for corruption, by checking whether increased illicit rents in the future
can reduce corruption today. Taking advantage of a change in wages in the
NREGS in Odisha that increased the possibility of higher rents to corrupt
officials in the future, they find that corruption indeed was significantly
lower today in areas that expected more lucrative projects to be forthcom-
ing. While on the one hand this provides support for the efficiency wage
hypothesis, the magnitudes of wage increases necessary to reduce corruption
are out of the range of what might be deemed feasible: they find that corrupt
rents were 100 to 1,100 times official wages.
A corollary to providing disincentives to engaging in corruption is to
provide incentives to uncover corruption, by both protecting and incentiv-
izing whistleblowers. The Whistleblowers Protection Act (which came into
force in May 2014), however, appears to fall short in this regard. In balancing
the rights of potential whistleblowers against the rights of officials to do
their jobs free of harassment, the current provisions seem to tilt the balance
towards officials. For example, there is no penalty for victimization of com-
plainants, the Central Vigilance Commission (charged with investigating
complaints) has no power to impose penalties, and—unlike the comparable
US law, for example—there is no provision for the whistleblower to be
compensated from any funds recovered as a result of the complaint (PRS
Legislative Research 2011).
A final intervention in this category is the introduction of independent,
third-party auditing. Duflo et al. (2013) implement a field experiment in
Gujarat in which they introduce the random assignment of auditing in two
heavily industrial regions of the state with the universe of “audit-eligible”
plants. The intervention has several components. First, the researchers
randomly assigned a third-party auditor to treatment plants who were paid
from a central pool rather than by the firm directly. Second, the researchers
verified a random sample of each auditor’s pollution readings with follow-up
visits. Finally, halfway through the study, treatment auditors were told that
their pay would be linked to the accuracy of their audit reporting.
The treatment resulted in more truthful reports by auditors for treated
firms. In terms of actual pollution outcomes, the treatment also succeeded
in reducing emissions with the “dirtiest” plants reducing emissions the most.
Finally, the authors demonstrate nonexperimental evidence that the financial
incentives for reporting accuracy, implemented mid-way through the study,
had an independent positive effect on reporting quality.
Hence, the introduction of independent “third-party” auditing resulted
in more accurate reporting, less “cheating,” and improved environmental
outcomes. The random assignment of independent auditors was clearly a
key component of the intervention, but equally important is the fact that
the auditor was paid from a central pool of funds rather than directly by the
audited firm. One could imagine a scenario in which the two are not de-
linked, which could potentially undermine the benefits of a third-party audit.
4.4. Electoral Reform/Political Incentives

When it comes to electoral reform, there are two distinct pathways through
which policymakers seeking to curb the influence of corrupt or criminal
politicians can operate. The first is to address the supply of “tainted” can-
didates into the electoral domain, while the second is to tackle the demand-
side incentives.
To curtail the entry of candidates associated with illegal activity in the

electoral domain, the Supreme Court issued two important judgments in 2013.
The first found that any MP or MLA currently holding office, if convicted
by a court of law (for charges listed under Sections 8(1), 8(2), or 8(3) of the
Representation of the People Act, 1951) would be immediately disqualified
from the date of conviction (unless he or she obtained a stay on the convic-
tion). Prior to Supreme Court intervention, convicted lawmakers could hold
on to their seat as long as an appeal on that conviction was pending before
the courts.17
The Supreme Court’s order disqualifying convicted legislators has
already had some impact. Shortly after the judgment, Bihar MP (and for-
mer Chief Minister) Lalu Prasad Yadav was forced to give up his seat in
Parliament after the prosecution obtained a conviction in the fodder scam
case. More recently, the chief minister of Tamil Nadu, Jayalalithaa, was for-
cibly removed from office after being convicted on graft charges, although
her conviction was later thrown out by an appeals court. According to an
analysis by ADR, there are 53 members of the 16th Lok Sabha who are at
risk of being disqualified and, hence, vacating their seat should pending
cases against them result in a conviction (and should they fail to obtain a
stay on that conviction). However, converting charges framed by a court
into a conviction is no small task. Of the 76 MPs serving in the 15th Lok
Sabha who faced serious ongoing criminal action, the average case faced
by this group had been pending for seven years (with some cases pending
for 25 years or more).
The court also ruled, in a separate judgment, that any candidate who was
either in jail or in police custody could not stand for elections (on the logic
that such a candidate is not an eligible elector in the election) under Indian
Law. Political parties across the spectrum alleged that the ruling provided
perverse incentives for the government to falsely imprison or detain political
opponents. In response, parliament swiftly passed a bill that clarified that that
a person does not cease to be an elector even if he or she cannot vote due to
police custody or incarceration. This effectively nullified the court’s ruling.
In addition to judicial action on removing convicted legislators, the
Election Commission and many good governance campaigners have
also suggested preventive action. For instance, the ECI has formally
17. In order to counteract the court’s move, the UPA government quickly introduced a bill
in parliament that would nullify the ruling, but it was later forced to withdraw the bill under
harsh public criticism. The government also considered promulgating an ordinance to the
same effect, but again had to retreat in light of public outrage.
recommended that candidates against whom charges have been framed by

a Court should be disqualified from contesting elections. In order to guard
against politically motivated charges, the ECI has suggested two additional
caveats: that only cases filed at least six months prior to the election and
those involving offences punishable by imprisonment of five years or more
should be considered (on the premise that cases which meet these criteria
are less likely to be subject to political motivation). Odisha MP Jay Panda
has introduced a private member bill which, if passed, would subject such
politicians to a newly created “fast track” tribunal to handle cases facing
politicians. Recently, Prime Minister Modi rhetorically backed the idea,
requesting that the Supreme Court set up a judicial mechanism to expedite
cases against “any Members of Parliament upon whom an FIR is lodged,”
and reach a verdict within one year.
There have also been numerous steps either initiated or proposed to
crack down on “black” money in politics (see Gowda and Sridharan 2012;
Sridharan 2006 for a review). There is, as we discussed above, a well-
established link between illicit election finance and criminality in politics,
or what is referred to in common parlance as “money and muscle.”
One basic remedy would be to improve the transparency of political party
finances, which is sorely lacking. To this end, in June 2013, the Central
Information Commission (CIC) ruled that for the purposes of the Right to
Information Act, political parties were to be considered “public authorities”
and, as a consequence, subject to provisions of the Act. Parliament quickly
introduced a bill to remove political parties from the ambit of the RTI Act,
but the bill was not passed during the 15th Lok Sabha. The CIC’s ruling is
controversial even for proponents of enhanced transparency, who contend
that it is the ECI, not the CIC or the RTI, which should have jurisdiction over
the affairs of political parties. However, in March 2015, the issue became
moot as the CIC effectively threw up its hands, declaring that its decision
was not implementable since parties had simply refused to cooperate with
the agency and it had no mechanisms by which to compel the participation
of the parties (Shrinivasan 2015).
Going forward, the ECI has proposed that it be granted greater author-
ity to regulate political parties. For instance, the ECI has no ability under
existing law to de-register political parties who flout democratic norms or
who set up parties with the sole purpose of exploiting tax loopholes. One
longstanding suggestion, also supported by many civil society organizations,
is that political parties submit annual, independent audits of their finances to
the ECI for public dissemination. To date, parties have resisted this move.
A second remedy under consideration is to enhance the ECI’s authorities

with respect to election-related disclosures. For instance, the ECI has also
been in a pitched legal battle with the Government of India to bring charges
against candidates who knowingly file false affidavits detailing election
expenditures. The UPA-2 Government openly challenged the commission’s
power to disqualify a candidate for falsifying election finance filings, stat-
ing the Commission has the power to sanction candidates who do not file
disclosures but not necessarily those who file incorrect ones (Hindu 2013).
In a closely watched case in July 2014, the ECI framed charges against
Congress MP Ashok Chavan, the former Maharashtra Chief Minister, for
filing false election expenditure disclosures, accusing him of willfully hid-
ing the fact that he paid off journalists to write positive news articles about
him during a 2009 election campaign. The proceeding was seen as a test
case of the agency’s powers, but in September 2014, the Delhi High Court
exonerated Chavan in the paid news case before the announcement of the
state assembly elections in Maharashtra (Garg 2014). In order to avoid future
legal uncertainty, the ECI has asked that the law be amended to clarify the
ECI’s authorities to sanction those who file false disclosures. In a similar
vein, the ECI has suggested that “paid news,” the practice of politicians
paying journalists for favorable media coverage masquerading as ”news,”
be explicitly classified as an “electoral offence” and made punishable under
the Representation of the People Act.
All of these proposed reforms deal with reducing the supply of poten-
tially “tainted” politicians into electoral politics. They work by changing
the incentives of political parties to give tickets to such politicians, reduc-
ing the flow of illicit money in politics, or simply disqualifying politicians
from contesting elections or holding office. Yet, there is a “demand” side
to the equation as well, which relates to voter incentives and the electoral
popularity of allegedly criminal or corrupt candidates. None of the remedies
described above on the supply side grapple with the basic notion that, at the
end of the day, “tainted” legislators would not be in office were it not for
voters electing them.
When it comes to voter motivations and support for criminal candi-
dates, there are broadly two schools of thought. The first believes this is an
issue about information asymmetries and “ignorant voters.” The second,
in contrast, believes that the appeal of criminal candidates is related to
their credibility to act as effective representatives. According to this view,
voters support such criminal candidates because, rather than in spite, of
their criminality.
There are a few studies that have explicitly tested the proposition that
improving the awareness of voters through the provision of information
reduces support for criminality. Banerjee, Kumar, Pande, and Su (2011)
conducted an experiment with Delhi slum dwellers in advance of municipal
elections in which a local NGO distributed newspapers containing report
cards on politicians to randomly selected residents. The report card pre-
sented information on the performance of the incumbent legislator and the
qualifications of the incumbent and two main challengers. The report card
contained information on legislative activism, legislator performance, and
expenditures from the incumbent’s local constituency development fund.
In addition, the report card contained data, gleaned from candidate affida-
vits, on educational qualifications, criminal records, and financial assets.
Relative to control slums, the researchers find that treatment slums (which
received the report card) experienced higher voter turnout, reduced vote
buying (measured directly through participant observation), and increased
support for better performing and more qualified incumbents. Interestingly,
however, information on criminality seems to have no impact. The results
indicate that neither information on the criminality of incumbents nor on
that of challengers have any statistically significant impacts on incumbent
vote share.
A paper by Banerjee, Green, Green, and Pande (2010b) reports on a voter
mobilization, as opposed to information, campaign in Uttar Pradesh. In a
set of randomly selected villages, an NGO conducted meetings and puppet
shows to mobilize voters. In the first treatment, the NGO urged people in
the village to vote on Election Day but to do so responsibly by voting on
issues, not on caste. The second treatment involved an abstract plea, by
the same NGO, to vote for “clean politicians.” The treatment imparted the
message: “Corrupt politicians steal money set aside for development funds
and do nothing for you. Vote for clean politicians that care about your
development needs.”
The results of the first treatment indicate that it both has a positive impact
on raising voter turnout and also reduces the extent of caste-based voting
(as measured by a follow-on survey). The reduction in caste-based voting is
linked to a significant decline in support for candidates charged with heinous
crimes (although not those deemed to be “corrupt,” as measured by a survey
of local journalists). The authors conclude that low-information voters who
are urged not to vote on the basis of caste will consider alterative evaluative
criteria (such as past criminality). They surmise that support for criminal
candidates, then, is a by-product of caste-based voting.
The null result on the corruption measure, however, complicates the

picture. The second treatment, urging voters not to elect corrupt candidates,
had no impact on voter turnout or vote share of corrupt or criminal candi-
dates. The authors reconcile these divergent findings by arguing that the
nonpartisan anticorruption campaign may have been too abstract and failed
to provide sufficient information needed to reshape voter preferences. On
the other hand, the anti-caste/pro-development messaging directly addressed
an important voter “heuristic,” while simultaneously offering voters a new
evaluation criterion (i.e., development).
It should be emphasized that the latter study relies on a voter mobilization,
rather than voter information and treatment; and the two are not identical.
Voter information had no impact on support for criminal candidates while
the anti-caste voter mobilization campaign did have an impact. This latter
finding is in line with Chauchard’s (2013) work, which finds that support
for criminality is not necessarily support for “criminals” per se, but a by-
product of ethnic voting.
The second school of thought regarding the success of criminal politi-
cians treats the issue of criminality in politics not as an information problem
but one of credibility. Criminality in this case, researchers have shown, is
often expressed through the language and symbolism of identity politics.
This is consistent with a growing qualitative and ethnographic literature on
criminality in Indian politics.
Recent survey data analyzed by Sircar and Vaishnav (2015) finds that
at least 26 percent of respondents from an all-India sample admitted that
they would vote for a candidate with serious criminal charges but who also
delivers benefits to them. What is especially intriguing is that there is a
strong association—at the state level—between support for criminality and
expressions of caste bias (as measured through a list experiment from the
same survey). This finding is in line with the “criminality as credibility”
hypothesis.
Thus, the prospects for addressing the demand for criminal politicians
are mixed. On the one hand, Banerjee et al. (2011) find that updating voter
information on the criminal records of incumbents and challengers has no
impact on voter behavior in Delhi. This is consistent with the literature that
downplays the role of information; perhaps telling voters about criminal
antecedents is not actually updating their beliefs in a meaningful way. Yet
Banerjee et al. (2010b), in their study from Uttar Pradesh, seem to suggest
that while voters may not be susceptible to the provision of factual infor-
mation on candidate criminality, they might be influenced by a hortatory
voter mobilization campaign. The literature reveals that there is clearly a
linkage between identity politics and support for criminality; what remains
unresolved is whether this connection is incidental or represents something
deeper. In other words, do voters support candidates who are co-ethnics
and hence deemed more credible (in which case, criminality is incidental)
or is the credibility of co-ethnicity in some way conditioned by criminal-
ity? Clearly, more research is needed on the “demand” side before we can
disaggregate these nuanced relationships.
4.5. Legal Reform

This category of anticorruption remedies involves enacting new laws to curb
corrupt activities. There is, indeed, a large legislative agenda—comprised
of several bills already introduced in parliament—on this score.
As mentioned previously, the Right of Citizens for Time Bound Delivery
of Goods and Services and Grievance Redressal Bill (2011) seeks to create
a mechanism to ensure the timely delivery of publicly provided goods and
services to citizens. The Bill requires all public authorities to appoint officers
to redress grievances and, if grievances are not redressed within 30 days,
financial penalties are imposed on the relevant bureaucrat. Many states have
set up identical mechanisms, raising questions about duplication as well as
jurisdictional authority given that many goods and services are state subjects
and involve state-level bureaucrats (PRS Legislative Research 2012). The
central bill, much like its various state-level incarnations, is meant to address
acts belonging to the first category of corruption mentioned in Section II,
“facilitative corruption.”
The Public Procurement Bill (2012) could help curb abuses in the cat-
egory we call “collusive corruption,” such as kickbacks from government
procurement, by regulating central government procurement and improving
its transparency. The bill establishes an open competitive bidding process
as the default for public procurement (unless otherwise justified). The bill
also mandates the publication of all procurement-related information on a
central portal.
The Public Procurement Bill, as well as the Prevention of Corruption
(Amendment) Bill (2013) and the Prevention of Bribery of Foreign Public
Officials and Officials of Public International Organizations Bill (2011), all
contain provisions on bribery. For example, the Procurement Bill criminal-
izes the acceptance of a bribe by a public servant as well as the offering of a
bribe or exerting “undue influence” on the procurement process by prospec-
tive bidders. The Prevention of Corruption Bill criminalizes the act of bribe-
giving and expands the definition of bribe-taking (which is already illegal
under Indian law). The latter bill makes giving bribes to, or receiving bribes
from, foreign public officials or international public officials a criminal act
(PRS Legislative Research 2014). This bill, if passed, would also ratify the
United Nations Convention Against Corruption, which India signed nearly
one decade ago in 2005. This highlights the role that external anchoring,
whereby membership in an international body incentivizes domestic reform,
can play in India’s fight against corruption (Morlino 2005).18
On the specifics of India’s bribery laws, the idea of criminalizing bribe-
giving, long part of the received wisdom, has recently been contested. In a
well-known paper, Basu (2011) argues that, insofar as “extractive corrup-
tion” is concerned (or harassment bribes ordinary citizens must pay in order
to receive what they are legally entitled to), the act of bribe giving should
be deemed “legitimate.” Basu argues that this will result in a sharp decline
in the incidence of bribe giving, and the reasoning is actually quite simple.
To quote the author: “[O]nce the law is altered in this manner, after the act
of bribery is committed, the interests of the bribe-giver and the bribe-taker
will be at divergence. The bribe-giver will be willing to cooperate in getting
the bribe-taker caught. Knowing that this will happen, the bribe-taker will
be deterred from taking a bribe.”
A paper by Abbink, Dasgupta, Gangadharan, and Jain (2014) experimen-
tally tested Basu’s theoretical proposal, using a lab experiment conducted
with university students in India. The authors find that when the bribe-giver
is provided legal immunity, reporting of bribe demands increases while the
demand for bribes declines (compared to a scenario in which both bribe-
giver and bribe-taker are held liable). This core finding is consistent with
Basu’s theoretical predictions. The authors, however, extend the analysis in
two important ways. They authors test how bribe-givers react when bribe-
takers can retaliate and when the bribe-giver’s payment is not refunded
(hence eliminating monetary incentives). Their findings suggest that strict
financial incentives do not overwhelmingly influence reporting behavior.
However, when retaliation is an option, bribe demands and reporting are
roughly on par with the baseline case (when both bribe-givers and takers
are legally liable). The authors thus conclude that asymmetric liability alone
18. Similarly, India’s decision to join the Financial Action Task Force (FATF), an inter-
national collective of nations dedicated to combating money laundering and terrorist finance,
helped to galvanize a series of reforms to modernize India’s financial and regulatory regime
in order curb the flow of illicit funds. These alterations, many of which were contained in The
Prevention of Money Laundering (Amendment) Bill, 2009, were stipulated as pre-conditions
to India’s joining the group as a full member in 2010.
may not reduce corruption. Legitimizing bribe giving must proceed hand in
hand with implementing procedures to limit retaliation.
In addition to legislation specifically designed to take action against
corruption, there are also legal changes that structure (or restructure) eco-
nomic interactions in ways that might minimize corruption. For instance,
in the lucrative domain of natural resources, the NDA government has
already taken steps to enact legal changes that will act to curtail collusive
corruption. During the decade of the 2000s, on account of skyrocketing
commodity prices, growing Chinese demand, and the booming domestic
economy, India’s natural resource sector was rife with rent-seeking. One
of the gravest abuses was the arbitrary and wholly discretionary power of
government authorities to allocate resources (such as spectrum, mining
licenses, and land) to hand-selected proprietors. The discretion inherent in
allocation rules paved the way for several of the biggest corruption scandals
of the 2000s. Although the government’s hand was forced by series of tough
Supreme Court judgements, in March 2015 the NDA successful obtained
parliamentary approval for two reform bills: the Coal Mines (Special
Provisions) Bill (2015) and the Mines and Minerals (Development and
Regulation) Amendment Bill (2015). Both bills compel the government to
allocate mining leases via auction.19 As scholars have argued and history has
confirmed, auctions can be an improvement but in and of themselves are no
panacea (Klemperer 2002). Indeed, in March 2015 media reports surfaced
suggesting the Government might cancel newly auctioned coal licenses on
suspicion that private firms may have colluded in the auction process to
artificially lower bid prices (Economic Times 2015).
4.6. Policy Reform

The final category of strategies to combat corruption involves simple reforms
in rules and regulations. Under this heading, we discuss user fees, streamlin-
ing of permissions, and the rule of law.
A straightforward fix to many types of day-to-day corruption, particularly
the type described in Section 2.1, is to institute user fees for faster delivery
of public services. This would simply take the place of “speed money”
bribes, hence redistributing these resources from wayward officials to gov-
ernment coffers, while at the same time reducing inefficiency by removing
uncertainty and not taxing honest citizens who may refuse to pay bribes.
19. The one exception relates to the 204 coal leases cancelled by the Supreme Court in
August 2014 on the heels of the ”Coalgate” scandal. The government can provide licenses
for these coalmines either through allotment or auction.
Such systems are common all around the world, and have been successfully
implemented in India through the implementation of “tatkal” (immediate)
schemes for obtaining train tickets, passports, phone services, etc. As one
former academic who is now a senior member of government noted tatkal
has basically eliminated bribes in obtaining railway tickets.
Another example of low-hanging fruit when it comes to policy reform
is the streamlining of permission to minimize the prevalence of harassment
bribes. In October 2014, Prime Minister Narendra Modi announced the
creation of a new online portal that would allow more than 600,000 firms
doing business in India to essentially “self-certify” their compliance with 16
different labor laws currently in force (Yadav 2014). This is a measure many
leading business groups had been clamoring for, and a step some investor-
friendly states have already pursued (FICCI 2014). Under the status quo,
each of the 16 laws provided an opportunity for government inspectors to
harass firms and seek bribes in exchange for filing favorable compliance
reports. Under the new system, the feared “Inspector Raj” will not com-
pletely disappear, but instead be replaced by a system of random audits to
ensure that participating firms are truthfully reporting compliance. While
the new regime allows for some discretion on the part of the government,
it significantly constrains it.
A final area deserving of greater attention is the rule of law. Intellectually,
the rule of law is a slippery concept to pin down because there is, of course,
no single institution that is charged with protecting the “rule of law.” As
Kapur and Vaishnav (2014) have argued, it is more useful to conceive of
the rule of law as a set of linked activities in a supply chain. On one end,
there are the laws on the books and the lawmakers who write them. Further
along the chain come the courts, which are in charge of adjudicating the
laws. At the opposite end sit prosecutors and the police, who are the front-
line functionaries of the state in charge of enforcement. Weaknesses or
shortcomings in any single link have obvious repercussions for the entire
chain. For instance, what good are sterling laws on paper if the police do
not have the capacity to enforce them?
Without reforming the rule of law in India, it will be impossible to
deter corrupt acts before they take place or to take action once corrupt acts
eventually come to surface. Any strategy to reform the rule of law must
begin with reforming India’s legal undergirding. In the previous section,
we discussed several new laws that could have a salutary impact on the
corruption environment. But just as important as adding new laws on the
books is eliminating outmoded and outdated ones. The regulation of labor,
for instance, is a domain badly encumbered with onerous and excessive
laws that do more to provide venal government officers with tools to extort
businesses than to protect the rights of workers.
With regards to the remaining links in the chain, under-capacity is a major
constraint. As we discussed in Section 2, India’s rule of law institutions are
woefully undermanned in personnel terms. Even leaving aside the possi-
bility that contractors or private sector employees might be able to correct
for an under-provision of public sector workers (think, for instance, of the
booming labor market for private armed guards), the strength of most law
enforcement entities is well below where it should be. Even after boosting
alternative justice mechanisms or systems of grievance redressal, there are
limits to the progress that can be achieved short of hiring more judges. One
solution which has been recommended by the Law Commission of India is
to create an all-India judicial service, akin to the other all-India civil services
(like the IAS). This idea was first mooted in the 1960s, and was endorsed
as recently as 2012 by a Committee of Secretaries chaired by the Cabinet
Secretary, but has yet to see the light of day due to continuing differences
among the various state governments and high courts (Times of India 2015).
Even if political disagreements could be overcome, the crunch of finan-
cial resources poses a considerable obstacle to boosting the capacity of the
public sector. While resources are a major consideration, one should not
overestimate their role as the binding constraint. Recent experimental work
by Banerjee, Hanna, and Mullainathan (2012) on police reform in Rajasthan
demonstrates how relatively costless policy changes can greatly aid the
capacity of existing public sector agencies. Using a randomized design, the
researchers implemented a series of recommendations repeatedly suggested
by various police reform panels over the years, including limiting arbitrary
transfers, enforcing the rotation of duty days and days off, increasing com-
munity involvement and training, and deploying “decoy” visits. Due to the
autonomy of middle managers, many of the interventions were not effec-
tively implemented. However, the researchers did detect significant impacts
from the training and decoy interventions.
Of course, a final impediment to reform is political will. In many
instances, politicians prefer the status quo to pushing for reform because
it protects their ability to use (and abuse) their discretionary authority to
benefit themselves. Consider the consequences for the main parties involved
in recent high-profile corruption scandals. On the one hand, it appears as
though the state is fairly good at taking obvious action upon discovery of
the scam: canceling contracts or licenses, starting investigations, and even
arresting the main parties involved. True, many of these actions reflect the
will of the courts rather than incumbent politicians, but nonetheless some
action is taken.
When the dust settles, however, the news is quite depressing. Referring
back to our inventory of corruption scandals since the year 2000, the most
powerful actors—defined as serving or having served at the level of Chief
Minister of a state or above—who are implicated in alleged wrongdoing
are never actually found guilty (none out of the 12 cases in our sample).
Instead, the pattern is to arrest involved parties and file a CBI case as soon
as the scam breaks, but following up on this is practically impossible in the
cases of such big players, and each of them is out on bail.
On the other hand, those with limited political protection, particularly
those involved in financial scams defrauding other parties, are much more
likely to actually serve jail time, as the experience of Ketan Parekh, Abdul
Kareem Telgi, and even Harshad Mehta in the past suggest. The recent
conviction of two former chief ministers (Lalu Prasad Yadav in Bihar, Om
Prakash Chautala in Haryana) and one sitting chief minister (Jayalalithaa
in Tamil Nadu) presents a hopeful sign that the culture of impunity may be
coming to an end, but it is too early to declare a new trend.
It is precisely because of weaknesses of existing agencies tasked with
combating corruption that good government campaigners lobbied for
Parliament to create a Lokpal, or anticorruption ombudsman with enhanced
powers to go after malfeasance in government. The bill, which was passed
in December 2013—45 years after it was first discussed on the floor of
Parliament—also mandates that states establish state-level ombudsmen,
or Lokayuktas, to curb corruption at the subnational level. Even before
Parliament passed the bill, at least 18 states already had such agencies up
and running.
It is too soon to tell whether the Lokpal will be more effective than India’s
existing graft-fighting institutions because, at the time of writing, the agency
is still in the process of being established and the government is still framing
the rules that will guide the new agency. However, there are a number of
contentious issues, yet to be resolved, that will determine the new body’s
effectiveness (see also Kapur and Vaishnav 2014).
After nearly one year in office, the NDA government has not yet initi-
ated the process of appointing a chairperson and the various subordinate
members who will manage the new agency. In addition, there are numerous
procedural issues regarding the Lokpal’s investigative powers that have not
been finalized. For instance, there is no clarity regarding how the Lokpal
will receive complaints from those who are affected by, or who witness,
corrupt acts. If the new agency requires approval by state governments in
order to investigate central government officials on state deputation, as is

widely understood to be the case, the end result will be a continued lack of
a unified body to decide all cases of sanction for prosecution. Furthermore,
while the Lokpal Act gives the new agency broad powers of superintend-
ence over the CBI, how the lines of authority actually operate in practice
remains to be seen.
5. Political Economy of Reform
In this penultimate section, we reflect on some of the deeper political

economy considerations related to fighting corruption in India. We begin
by addressing the question of the political conditions under which anticor-
ruption reforms take place. We then discuss the thorny issue of relying on
politicians associated with criminality and corruption conduct to actually
implement anticorruption reforms. We end by discussing new research
on voter behavior and whether shifts among the electorate might alter the
prospects of politicians linked with corruption.
5.1. Conditions for Reform

We begin by documenting some of the most frequent conditions under
which anticorruption reforms gain traction. Our core motivation here is to
get a better handle on the circumstances under which political space can be
created in democratic societies to enact serious anticorruption remedies. To
what extent, for instance, can India learn from the example of other societies
that have evolved out of patronage politics and toward a pro-development
equilibrium?
5.1.1. POLITICAL RETRIBUTION

The most obvious response to the question of when political space arises
to address corruption is when the incumbent can use the issue to attack
its political opponents. Unfortunately, such an approach is likely to fail
because it is opportunistic and reeks of arbitrary enforcement or manipula-
tion. Consider, for instance, the farcical case of former Uttar Pradesh chief
minister and Samajwadi Party (SP) President Mulayam Singh Yadav, who
the CBI first charged with corruption (in Indian parlance, “accumulating
disproportionate assets”) in March 2007 (Srivastava 2013). In July 2008,
the SP provided outside support to the UPA government in New Delhi when
it was facing a vote of no confidence over the contentious US–India civil
nuclear bill. Months after the SP bailed out the UPA government, the CBI
withdrew the case against Yadav, claiming it possessed insufficient material
to prosecute. Yet a little more than two years later, in the spring of 2011,
the SP and the Congress Party were at loggerheads over a potential seat-
sharing agreement in the coming Uttar Pradesh assembly elections. Months
later, the CBI reversed its stand and reintroduced charges against Yadav. In
December 2012, the SP once again came to the aid of the Congress central
government by allowing a controversial bill involving foreign direct invest-
ment in multi-brand retail to sail through Parliament. By September 2013,
the CBI had once more closed its case against Yadav, citing “grossly insuf-
ficient evidence.” Clearly, using the state machinery to prosecute alleged
corrupt acts on the basis of political motivations has inherent shortcomings.
5.1.2. ACCESS TO ALTERNATIVE SOURCES OF RENTS

A second condition under which political elites can take on corrupt activities
is when they have access to other financial flows that can compensate for
foregone rents. For instance, one hypothesis about why retail corruption has
declined in states such as AP and Tamil Nadu compared to Bihar is that in
the former states, the rents politicians (and bureaucrats) used to extract from
service delivery or entitlement programs have been replaced with rents (often
larger in magnitude and less cumbersome to collect) from infrastructure or
contracts. While such hypotheses are difficult to test, this view is backed
up by multiple conversations the authors have had with senior IAS officers,
politicians, and academics.
Moreover, it is also possible that at least some of these big-ticket rents
have less of an impact on the economy than regular corruption. For exam-
ple, it is apparently more and more common to pay officials and politicians
via equity stakes rather than bribes. This type of quid pro quo more closely
aligns officials’ and businessmen’s incentives in support of the eventual suc-
cess of the project. In their work on political finance, Kapur and Vaishnav
(2015) provide one such case: “a builder constructing a hotel in Mumbai told
the authors that the government told him it would only issue building permits
if there was a quid pro quo. The quid pro quo sought was not cash but a
5 percent equity stake in the hotel in the name of a firm connected to a local
politician.”20 In other interviews conducted by one of us (Vaishnav), several
builders in the state of Gujarat claimed that it was common knowledge that a
senior level politician insisted on a 5 percent equity stake in any new major
20. Kapur and Vaishnav cite an interview with a builder in New Delhi on December 3, 2012.
construction project in the city of Rajkot, where his constituency is located

(Author interview, Ahmedabad, December 2012).
Academic evidence, albeit limited to one article, also supports this view.
The much-publicized 2G scam apparently had no effect on wireless telecom
markets, basically because the illegally acquired licenses were then sold off
to legitimate telecom operators at a large cost to the government but not
necessarily at the cost of consumers (Sukhtankar 2015).
5.1.3. POPULAR ANTICORRUPTION MOVEMENTS

Another condition under which anticorruption reforms can come to the
fore is when strong movements create political space for reformers to push
through pivotal changes. This is what transpired in the decades following
the Gilded Age in the USA, one popular historical example of political cor-
ruption, as the country transitioned toward what later became known as the
“Progressive Era,” a period spanning from roughly 1890 to 1920.21 During
this three decade-long stretch, the USA saw a number of major reforms to
the structure of American government and society, underpinned by a new
conception of the state’s role in private affairs. The Progressive Era was a
product not of any single reform movement but at least four separate ones:
the “business regulation” movement, which fought to introduce competition
in industries that had reaped the benefits of consolidation and monopoly;
the “good government” movement, which tackled the corrupt machines
that held sway during the Gilded Age and sought to introduce new manage-
rial practices into governance; the “social justice” movement, which was
motivated by upgrading the quality of life for the urban poor (especially
immigrants and factory workers); and, finally, the “social control” move-
ment, which was imbued with a moral imperative to inculcate middle-class
Christian values into society.
In India, reformist anticorruption movements have been episodic to date.
The most recent example is the India Against Corruption (IAC) movement,
launched by Anna Hazare, Arvind Kejriwal, and their colleagues and sup-
porters, which shot to prominence in 2010 following revelations of massive
corruption scandals in the final years of the UPA-2 government. IAC was
narrowly focused on pressuring the government to enact what it called the
“Jan Lokpal” bill, which was more sweeping that the government’s draft
Lokpal bill. Although the government eventually passed a revised version
21. The authors are grateful to Alec Sugarman for his research on the American historical
case.
of its original bill in 2013, IAC’s leadership fractured, with Kejriwal lead-
ing the faction that would eventually morph into a new political party, the
AAP. After its initial 49 days running the Delhi government, in which it
aggressively pushed anticorruption legislation, AAP’s majority government
in Delhi (formed in December 2014 following fresh elections) has been less
insistent. Whether AAP can make headway on corruption remains to be seen,
as is its ability to scale up to other geographies outside of the immediate
National Capital Region. Outside of winning four seats in Punjab, the party
performed poorly in the 2014 Lok Sabha election. To date, it has not suc-
ceeded in creating a pan-Indian organization that could expand its footprint
outside of Delhi. Furthermore, judging by the high degree of internal party
conflict it is experiencing at the time of writing, there is some evidence that
the party is possibly sacrificing its fervent anticorruption modus operandi
for political expediency (Joshi 2015). There have been similar instances in
India’s post-Independence history of movements coming to the fore (the JP
movement in the 1970s or the example of V.P. Singh in the 1980s) motivated
by an underlying obsession with fighting corruption only to quickly dissi-
pate or be overtaken by other issues once they entered the political domain.
By and large, IAC, AAP, and other such groups are in line with the “good
government” strand of the Progressive Era reform movement. India has also
seen elements of the “social justice” movement discernible in many of the
rights-based movements of the 2000s around issues such as the right food,
water, employment, shelter, etc. Unfortunately, these movements were far
more effective in advocating for new legislation to enshrine these rights than
in the actual design of social programs that emerged out of this rights-based
agenda. Notably weak is the “business regulation” movement, which in the
USA helped push landmark legislation such as the Interstate Commerce
Act of 1887, the Federal Trade Commission Act of 1914, and the Clayton
Anti-Trust Act of 1914. But while this strand of the movement is weak, it
is not totally absent. Many of the corruption scams we have documented
in Table 1 (6 of 28, or 21 percent) are from the financial services sector,
typically involving private parties or entities conning other private parties.
Indeed, there are likely direct correlations between the creaky regulatory
structure, outmoded laws and gaps in consumer protection, and corruption
in the financial sector. Partially in reaction to this, the Government estab-
lished the Financial Sector Legislative Reforms Commission (FSLRC) to
revamp India’s archaic legal and regulatory framework governing India’s
financial sector. The effort, described by Patnaik and Shah (2014), serves
as a potential template for future reform initiatives.
5.2. Voting for Corruption and Criminality

One difficulty in enacting anticorruption reforms is a practical one: many
politicians currently in power are suspected of engaging in criminal or cor-
rupt acts. As Section 4.4 has shown, there is a burgeoning literature on the
criminalization of politics in India.
Because “criminality” is a broad category, encompassing a range of
potential criminal activity, it is worth noting that descriptive analyses suggest
that the picture is not qualitatively different if one looks at a more narrow
definition of “corruption.”
To illustrate the point, drawing on the affidavits submitted by candidates
contesting the 2014 general election, we isolate those candidates who declare
pending criminal cases in which they are specifically charged with com-
mitting readily identifiable acts of “corruption” (Table 3). These charges
include violations of the Prevention of Corruption Act or any other cases
overtly related to vigilance or anticorruption charges, as identified by the
candidate on his or her affidavit. We identify 24 such candidates contesting
the 2014 general elections, representing a range of parties and a wide diver-
sity of states. Of these 24 candidates facing corruption cases, the median
candidate finished in second place, earning 37 percent of the constituency
vote. These candidates performed impressively, when considering that the
median candidate in the 2014 election finished in ninth place and earned
less than one percent of the vote. Indeed, nine of the “corruption-tainted”
candidates (37.5 percent) actually won their elections. When we examine
a smaller subset of seven candidates who had won in 2009 and stood for
reelection as incumbents in 2014, we see that corruption-accused candidates
endured only modest penalties for their alleged transgressions (Table 4).
Three of seven candidates claimed victory while five increased the number
of votes they won (indeed, 2G accused former Union Telecommunications
Minister A. Raja won nearly 42,000 more votes than 2009 even though he
lost his reelection) and the median vote share penalty was around 5 percent.
This data starkly portrays the dilemma facing India’s anticorruption
reformers: a significant proportion of the politicians they must rely on to
pursue reforms that might curb malfeasance are themselves suspected of
engaging in illegal activity. For obvious reasons, such politicians might
not be the most aggressive anticorruption reformers. But an even more
fundamental problem is that voters are affirmatively electing (and often
reelecting) such individuals.
In this vein, one encouraging development worth highlighting is the
significant shift underway in Indian voter behavior in recent years. The vast
TABLE 3. Electoral Performance of Candidates with Declared Corruption Cases, 2014 Lok Sabha Elections
Candidate Constituency State Party Rank Vote Share Elector Share
Captain Amarinder Singh Amritsar Punjab INC 1 0.48 0.33
Patil Nitin Suresh Aurangabad Maharashtra INC 2 0.37 0.23
B. Sreeramulu Bellary Karnataka BJP 1 0.51 0.36
Rakesh Dhar Tripathi Bhadohi Uttar Pradesh BSP 2 0.25 0.13
N. Dharam Singh Bidar Karnataka INC 2 0.38 0.23
Natubhai G. Patel Dadra & Nagar Haveli Dadra & Nagar Haveli BJP 1 0.49 0.41
Anbumani Ramadoss Dharmapuri Tamil Nadu PMK 1 0.43 0.34
Rakesh Sachan Fatehpur Uttar Pradesh SP 3 0.17 0.10
Anurag Singh Thakur Hamirpur Himachal Pradesh BJP 1 0.54 0.36
Agrawal Kunjbihari Jugalkishor Jalna Maharashtra SP 11 0.00 0.00
Prem Chand Vishavkarma Kangra Himachal Pradesh IND 7 0.01 0.00
Lalit Mohon Suklabaidya Karimganj Assam INC 3 0.26 0.19
Anosh Ekka Khunti Jharkhand JKP 2 0.24 0.16
Mohammed Faizal.P.P Lakshadweep Lakshadweep NCP 1 0.50 0.43
Paresh Baishya Mangalldoi Assam AIUDF 4 0.06 0.05
Ayodhya Rami Reddy Alla Narasaraopet Andhra Pradesh YSRC 2 0.47 0.39
Gokaraju Ganga Raju Narsapuram Andhra Pradesh BJP 1 0.50 0.41
A. Raja Niligiris Tamil Nadu DMK 2 0.38 0.28
Pongalur Palanisamy.N Pollachi Tamil Nadu DMK 3 0.25 0.18
Bandhu Tirkey Ranchi Jharkhand AITC 5 0.04 0.03
Ganesh Singh Satna Madhya Pradesh BJP 1 0.41 0.26
B.S. Yeddyurappa Shimoga Karnataka BJP 1 0.54 0.39
Atiq Ahmed Shrawasti Uttar Pradesh SP 2 0.27 0.15
Churchill Alemao South Goa Goa AITC 3 0.03 0.02
Bharat Ramaswami and Shikha Jha
Source: Author’s calculations based on affidavits submitted to Election Commission of India by candidates contesting the 2014 general election.
249
TABLE 4. Electoral Performance of Re-contesting Candidates with Declared Corruption Cases, 2014 Lok Sabha Elections
Change, Change, Change,
Candidate Constituency State Party Rank 2009 Rank 2014 # votes vote share elector share
Lalit Mohon Suklabaidya Karimganj Assam INC 1 3 –33155 –0.123 –0.049
Rakesh Sachan Fatehpur Uttar Pradesh SP 1 3 –39229 –0.145 –0.043

A. Raja Niligiris Tamil Nadu DMK 1 2 41958 –0.062 –0.033
N. Dharam Singh Bidar Karnataka INC 1 2 29111 –0.051 0.000
Anurag Singh Thakur Hamirpur Himachal Pradesh BJP 1 1 74437 0.002 0.044
Natubhai G. Patel Dadra & Nagar Haveli Dadra & Nagar Haveli BJP 1 1 29548 0.024 0.079
Ganesh Singh Satna Madhya Pradesh BJP 1 1 180664 0.116 0.096
Source: Author’s calculations based on affidavits submitted to Election Commission of India by candidates contesting the 2009 and 2014 general elections.
literature on Indian elections suggests that Indian voters have traditionally

conditioned their electoral choices on parochial issues such as ethnic-
ity, patronage, or clientelism, rather than issues concerning governance
and development (Chandra 2004). These priorities, in turn, have had an
adverse impact on the nature of political selection, providing the electoral
basis for politicians associated with criminal or corrupt backgrounds to
thrive (Banerjee and Pande 2009). There is some evidence to suggest that
the situation is gradually changing. For instance, a study by Gupta and
Panagariya (2014) finds that, contrary to the received wisdom, voters in
the 2009 Lok Sabha elections rewarded candidates associated with strong
economic records, with incumbents in “high-growth” states significantly
outperforming those from states that experienced either “moderate” or “low”
growth. In the 2014 general elections, post-poll survey evidence suggests
that economic issues, such as inflation, economic development, and lack of
employment, weighed heavily in voters’ minds as they cast their ballots on
election day (CSDS 2014).
The shift toward economic voting, which has been commonplace in many
advanced industrial democracies for many years, is discernible in state elec-
tions as well. Vaishnav and Swanson (2015) study the relationship between
economic growth and electoral returns in more than 120 state elections held
between 1980 and 2012. While they find little association between econom-
ics and elections in the 1980s and 1990s, they find a robust, positive relation-
ship in the post-2000s era. In the most recent period, the authors document
statistically significant electoral returns to those state governments able to
deliver faster rates of economic growth.
The rise of economic voting is encouraging from the perspective of
accountability, but the fact that economic voting coexists with a large preva-
lence of suspected criminal and/or corrupt elected representatives suggests
that there is not a one-to-one relationship between economic voting and a
decline in the salience of “parochial” interests. How do we square these
seemingly contradictory impulses—a shift toward retrospective economic
voting and a penchant for electing candidates with “tainted” backgrounds?
One possibility is that, in some circumstances, voters can simultaneously
vote on the basis of the macro-economy and for tainted politicians if their
vote choice is being driven by the identity of a chief ministerial or prime
ministerial candidate or a party label. Likewise, voters could have differ-
ing preferences for who they want to run the state or country as opposed to
represent them as an MLA or MP in their constituency.
A second, more troubling possibility is that voters are voting for better
governance, but tend to view tainted politicians in this light as well. In other
words, referring back to the notion that criminals often are deemed more
“credible” representatives, it could be that voters are selecting so-called
tainted politicians precisely because they view them to be effective repre-
sentatives in the context of a weak state. However, to the extent criminal or
corrupt politicians gain strength by filling a governance vacuum, they face
incentives to purse temporary, rather than lasting, solutions. If criminal rep-
resentatives, for instance, serve a useful role resolving disputes or providing
security, they have every reason to undermine lasting governance solutions
that risk making their services irrelevant.
6. Conclusion
The overall objective of this paper is to bring the research on corruption

and the policy discourse into the same conversation. Because the literature
on corruption in India is-to put it mildly-voluminous; our aim is to provide
a framework for thinking through the many issues raised by this body of
work. To that end, this paper offers a framework for thinking through the
underlying drivers of corruption in contemporary India, provides a rubric
for classifying corrupt acts, and discusses broad strategies for combating
corruption and what the literature tells us about their relative effectiveness.
To embed our overall discussion in India’s political context, we have also
offered some parting thoughts on India’s political economy of reform when
it comes to fighting corruption. To make our task manageable (and coher-
ent), we have made two important choices in this paper: to focus narrowly
on public sector corruption and to draw primarily from academic studies
that combine analytical rigor with causal analysis and rich sources of data.
In an effort to provide guidance to policymakers and other agents of change
addressing India’s corruption challenge, we conclude by highlighting five
principles that should guide future reform efforts.
First, information provision is an important tool in the toolbox, but, on
its own, it is not always an effective anti-graft strategy. Our review of the
literature reveals that information works best when accompanied by invest-
ments in enhancing the bargaining power of ordinary citizens, improving
coordination and collective action, or strengthening the state’s ability to
punish impunity. Social audits and information campaigns that uncover
malfeasance in a context of weak public sector enforcement institutions can
have limited impact, as demonstrated by Afridi and Iversen (2013) in their
study of social audits in Andhra Pradesh.
In the case of criminal or corrupt actors in electoral politics, there is com-

pelling evidence that the factors which give rise to this nexus are perhaps
less related to information asymmetries but instead have more to do with
social divisions embedded within India’s weak rule of law society. This is
not to say improving the information environment is not a laudable goal;
to the contrary, improving the availability of information on criminality
in politics has been essential to both diagnosing the challenge as well as
forging social pressure. The key take-away, however, is that the absence of
information may not the binding constraint.
Second, technological solutions to curb corruption have limited effective-
ness unless they are able to bypass the local machinery that hampers status
quo solutions. Technological innovations that require higher level authorities
to provide enforcement risk falling prey to the usual principal-agent dilem-
mas that plague public service delivery. Instead, interventions such as the
one involving smart cards in Andhra Pradesh evaluated by Muralidharan
et al. (2014a), which can transfer bargaining power to citizens and circum-
vent the broken local state machinery, hold significant promise. To this
end, the new NDA government in Delhi has an opportunity to build on the
Aadhaar program launched by the prior regime and further marginalize
middlemen in service delivery.
Third, there is a sensible and wide-ranging legislative agenda to reduce
corruption that the 16th Lok Sabha should pursue with renewed vigor.
Measures such as the “Right to Services” and the “Public Procurement”
bills contain important provisions that can constrain abuses of government
discretion while shifting bargaining power in favor ordinary citizens. To
be clear, these bills, as they stand, are imperfect; for instance, the “Right
to Services” bill would create a dedicated grievance redressal mechanism
in a context where there are legitimate concerns about the multiplicity of
grievance redressal mechanisms already in existence (on paper at least if
not always in practice). Parliament must debate and discuss these details
and forge reasonable compromise.
Reformers should also take heed that as meritorious as these bills may
be, passing news laws must be accompanied by a renewed effort to repeal
outdated or archaic old ones. Of course, it is natural for agents of change to
focus their attention on enacting new laws given the inherent benefits for
mobilization; yet such an approach is shortsighted. As Kapur and Vaishnav
(2014) argue, in India “the multiplicity and complexity of laws make com-
pliance, deterrence, and effective enforcement difficult and, in many cases,
impossible.” Statutes regarding corrupt practices are no exception.
Finally, while the state in India is often perceived to be the problem

when it comes to corruption, it is no doubt also part of the solution (Kapur
2010). There is a strong case to be made that the state, particularly the local
state, has historically preyed on the aam aadmi (common man), rather than
worked on its behalf (Pritchett 2009). Yet, there are limits to how much
can be achieved to reduce its corruption by circumventing, rather than
strengthening, its capacity.
At the end of the day, even the most immaculate laws require effective
state institutions to enforce them and judicial officers to adjudicate disputes.
Yet police vacancy rates in India hover around 25 percent and existing forces
are poorly trained, starved of resources, and the subject of constant political
interference. Similar shortcomings plague the judiciary at the same time
that the volume of litigation is rapidly increasing. The pioneering Right to
Information Act gives average Indians greater recourse to redressing griev-
ances than ever before, but if government information officers remain in
short supply and appeals processes drag on, empowerment could turn into
disenchantment.
The anticorruption agenda in India is massive but the enormity of the
task should not dampen the spirits of reformers. To be clear, the stakes are
high; this is not only because corruption can hamper India’s ability to grow
its economy and manage the enormous task of providing opportunities for
its young, growing population, but also because corruption can negatively
color popular perceptions of democracy and faith in the rule of law. Having
said that, reformers should take heed of the fact that the literature is replete
with successful examples of logistically simple solutions that can be imple-
mented at minimal cost. While the ability of these solutions to circumvent
weak public sector institutions has its limits, the potential gains from reform
suggest that such an agenda should be pursued with alacrity.
References
Abbink, Klaus, Utteeyo Dasgupta, Lata Gangadharan, and Tarun Jain. 2014. “Letting
the Briber Go Free: An Experiment on Mitigating Harassment Bribes,” Journal
of Public Economics, 111 (March): 17–28.
Acemoglu, Daron. 2010. “Theory, General Equilibrium, and Political Economy in
Development Economics,” Journal of Economic Perspectives, 24 (3): 17–32.
Afridi, Farzana and Vegard Iversen. 2013. “Social Audits and MGNREGA delivery:
Lessons from Andhra Pradesh,” in Barry Bosworth, Arvind Panagariya and
Shekhar Shah, eds, India Policy Forum 2013–2014. New Delhi: SAGE
Publications.
Agence France-Presse. 2015. “After 24-Year Leave, This Government Employee
Was Sacked Today,” January 8. Available online at http://www.ndtv.com/india-
news/after-24-year-leave-this-government-employee-was-sacked-today-724486
(accessed March 10, 2015).
Aidt, Toke, Miriam A. Golden, and Devesh Tiwari. 2013. “Incumbents and Criminals
in the Indian National Legislature,” Mimeo, University of California-Los Angeles.
Association for Democratic Reforms. 2014. “Analysis of Total Income of National
Parties between FY 2004–05 and 2011–12,” February 7. Available online at
http://adrindia.org/research-and-report/political-party-watch/combined-reports/
analysis-total-income-national-parties-be (accessed January 12, 2015).
Banerjee, Abhijit V. 1997. “A Theory of Misgovernance,” Quarterly Journal of
Economics, 112(4): 1289–1332.
Banerjee, Abhijit V., Rukmini Banerji, Esther Duflo, Rachel Glennerster, and
Stuti Khemani. 2010a. “Pitfalls of Participatory Programs: Evidence from a
Randomized Evaluation in Education in India,” American Economic Journal:
Economic Policy, 2(1): 1–30.
Banerjee, Abhijit V., Donald Green, Jennifer Green, and Rohini Pande. 2010b. “Can
Voters be Primed to Choose Better Legislators? Experimental Evidence from
Rural India,” Mimeo, Massachusetts Institute of Technology.
Banerjee, Abhijit, Esther Duflo, and Rachel Glennerster. 2008. “Putting a Band-Aid
on a Corpse: Incentives for Nurses in the Indian Public Health Care System,”
Journal of the European Economic Association, 6(2–3): 487–500.
Banerjee, Abhijit, Rema Hanna, and Sendhil Mullainathan. 2012. “Corruption,” in
Robert Gibbons and John Roberts, eds, Handbook of Organizational Economics.
Princeton, NJ: Princeton University Press.
Banerjee, Abhijit V. and Rohini Pande. 2009. “Parochial Politics: Ethnic Preferences
and Politician Corruption,” Mimeo, Harvard Kennedy School.
Banerjee, Abhijit V., Selvin Kumar, Rohini Pande, and Felix Su. 2011. “Do Informed
Voters Make Better Choices? Experimental Evidence from Urban India,” Mimeo,
Harvard Kennedy School.
Banerjee, Abhijit V., Donald Green, Jeffrey McManus, and Rohini Pande. 2012.
“Are Poor Voters Indifferent to Whether Elected Leaders are Criminal or Corrupt?
A Vignette Experiment in Rural India,” Mimeo, Massachusetts Institute of
Technology.
Bardhan, Pranab K. 1997. “Corruption and Development: A Review of Issues,”
Journal of Economic Literature, 35: 1320–46.
Basu, Kaushik. 2011. “Why, for a Class of Bribes, the Act of Giving a Bribe should
be Treated as Legal,” Mimeo, Finance Ministry, Government of India.
Becker, Gary S. and George J. Stigler. 1974. “Law Enforcement, Malfeasance, and
Compensation of Enforcers,” The Journal of Legal Studies, 3(1): 1–18.
Bertrand, Marianne, Rema Hanna, Simeon Djankov, and Sendhil Mullainathan. 2007.
“Obtaining a Driving License in India: An Experimental Approach to Studying
Corruption,” Quarterly Journal of Economics, 122(4): 1639–76.
Besley, Timothy, Rohini Pande, and Vijayendra Rao. 2007. “Just Rewards? Local
Politics and Public Resource Allocation in South India,” Mimeo, London School
of Economics and Political Science.
Bhavnani, Rikhil. 2012. “Using Asset Disclosures to Study Politicians’ Rents: An
Application to India,” Paper presented at World Bank Annual Bank Conference
on Development Economics, Washington, D.C., May 7–8.
Björkman, Martina and Jakob Svensson. 2009. “Power to the People: Evidence from
a Randomized Field Experiment on Community-Based Monitoring in Uganda,”
Quarterly Journal of Economics, 124(2): 735–69.
Bold, Tessa, Mwangi Kimenyi, Germano Mwabu, Alice Ng’ang’a, and Justin
Sandefur. 2013. “Interventions and Institutions: Experimental Evidence on
Scaling up Education Reforms in Kenya,” Working Paper 321, Center for Global
Development.
Bussell, Jennifer. 2012. Corruption and Reform in India: Public Services in the
Digital Age. New York: Cambridge University Press.
Centre for the Study of Developing Societies. 2014. India National Election Study
2014. CSDS, New Delhi.
Chandra, Kanchan. 2004. Why Ethnic Parties Succeed: Patronage and Ethnic
Headcounts in India. New York: Cambridge University Press.
Chauchard, Simon. 2013. “Why Do Voters Prefer Dynastic Candidates? Evidence
From a Vignette-Experiment,” Mimeo, Dartmouth College.
Chaudhury, Nanzmul, Jeffrey Hammer, Michael Kremer, Karthik Muralidharan, and
F. Halsey Rogers. 2006. “Missing in Action: Teacher and Health Worker Absence
in Developing Countries,” Journal of Economic Perspectives, 20(1): 91–116.
Cole, Shawn A. 2009. “Fixing Market Failures or Fixing Elections? Elections,
Banks, and Agricultural Lending in India,” American Economic Journal: Applied
Economics, 1(1): 219–50.
Das, Jishnu and Jeffrey Hammer. 2007. “Money for Nothing: The Dire Straits of
Medical Practice in Delhi, India,” Journal of Development Economics, 83(1):
1–36.
Duflo, Esther, Rema Hanna, and Stephen Ryan. 2012. “Incentives Work: Getting
Teachers to Come to School,” American Economic Review, 102(4):1241–78.
Duflo, Esther Michael Greenstone, Rohini Pande, and Nicholas Ryan. 2013. “Truth-
telling by Third-party Auditors and the Response of Polluting Firms: Experimental
Evidence from India,” Quarterly Journal of Economics, 128(4): 1499–1545.
Dutta, Bhaskar and Poonam Gupta. 2014. “How Indian Voters Respond to Candidates
with Criminal Charges,” Economic and Political Weekly, 49(4): 43–51.
Dhaliwal, Iqbal and Reema Hanna. 2014. “Deal with the Devil? The Success and
Limitations of Bureaucratic Reform in India,” Mimeo, Harvard Kennedy School.
Ferraz, Claudio and Frederico Finan. 2011. “Electoral Accountability and Corruption
in Local Governments: Evidence from Audit Reports,” American Economic
Review, 101(4): 1274–1311.
FICCI. 2014. Suggested Labour Policy Reforms. New Delhi: All India Organization
of Employers. Available online at http://www.ficci.com/SEdocument/20301/
FICCI-NOTE-ON-LABOUR-POLICY-REFORMS.pdf (accessed on March
15, 2015).
Fisman, Ray, Florian Schulz, and Vikrant Vig. 2014. “The Private Returns to Public
Office,” Journal of Political Economy, 122(4): 806–62.
Garg, Abhinav. 2014. “HC Dismisses Ashok Chavan ‘Paid News’ Case,” Times
of India, September 13. Available online at http://timesofindia.indiatimes.com/
india/HC-dismisses-Ashok-Chavan-paid-news-case/articleshow/42385479.cms
(accessed on November 20, 2014).
Gingerich, Daniel. 2014. “Brokered Politics in Brazil: An Empirical Analysis,”
Quarterly Journal of Political Science, 9(3): 269–300.
Golden, Miriam and Brian Min. 2013. “Distributive Politics Around the World,”
Annual Review of Political Science, 16: 73–99.
Gowda, MV., Rajeev, and E. Sridharan. 2012. “Reforming India’s Party Financing
and Election Expenditure Laws,” Election Law Journal, 11(2): 226–40.
Gupta, Poonam and Arvind Panagariya. 2014. “Growth and Election Outcomes in
a Developing Country,” Economics and Politics, 26(2): 332–54.
Harris, Gardiner. 2013. “India Aims to Keep Money for Poor Out of Others’
Pockets,” The New York Times, January 5. Available online at http://www.
nytimes.com/2013/01/06/world/asia/india-takes-aim-at-poverty-with-cash-
transfer-program.html?_r=0 (accessed on April 15, 2014).
Huntington, Samuel. 1968. Political Order in Changing Societies. New Haven:
Yale University Press.
Iyer, Lakshmi and Anandi Mani. 2012. “Traveling Agents: Political Change and
Bureaucratic Turnover in India,” The Review of Economics and Statistics, 94(3):
723–39.
Joshi, Sanjay. 2015. “The AAP and the 1935 Parallel,” The Hindu, April 2. Available
online at http://www.thehindu.com/opinion/lead/the-aap-and-the-1935-parallel/
article7058388.ece (accessed on April 5, 2015).
Kapur, Devesh. (2010). “Political Economy of the State,” in Niraja Gopal Jayal
and Pratap Bhanu Mehta, eds, The Oxford Companion to Politics in India. New
Delhi: Oxford University Press.
Kapur, Devesh and Milan Vaishnav. 2014. “Strengthening Rule of Law,” in Bibek
Debroy, Ashley J. Tellis and Reece Trevor (eds), Getting India Back on Track: An
Action Agenda for Reform, pp. 443–58. Washington, D.C.: Carnegie Endowment
for International Peace.
———. 2015. “Builders, Politicians, and Election Finance,” Mimeo, Carnegie
Endowment for International Peace.
Kenny, Charles and Jonathan Karver. 2012. “Publish What You Buy: The Case for
Routine Publication of Government Contracts,” Center for Global Development
Policy Paper, Washington, D.C.
Khera, Reetika. 2011. “Trends in Diversion of PDS Grain,” Technical Report
No. 198, Centre for Development Economics, Delhi School of Economics,
March 2011.
Klemperer, Paul. 2002. “What Really Matters in Auction Design,” Journal of
Economic Perspectives, 16(1): 169–189.
Krishnan, K.P. and T.V. Somanathan. Forthcoming. “The Civil Service in the 21st
Century,” in Devesh Kapur and Pratap Bhanu Mehta, eds, Building an Indian
State for the 21st Century.
Laffont, Jean-Jacques. 2006. “Corruption and Development,” in Abhijit V. Banerjee,
Roland Benabou, and Dilip Mookherjee, eds, Understanding Poverty. Oxford:
Oxford University Press.
Law Commission of India. 2014. “Arrears and Backlog: Creating Additional Judicial
(wo)manpower,” Report No. 245. Available online at http://lawcommissionofindia.
nic.in/reports/Report_No.245.pdf (accessed on January 17, 2015).
Leff, Nathaniel. 1964. “Economic Development through Bureaucratic Corruption,”
American Behavioural Scientist, 8(3): 8–14.
Lewis-Faupel, Sean, Yusuf Neggers, Benjamin A. Olken and Rohini Pande. 2013.
“Can Electronic Procurement Improve Infrastructure Provision? Evidence from
Public Works in India and Indonesia,” Mimeo, Harvard Kennedy School.
Lui, Francis. 1985. “An Equilibrium Queuing Model of Bribery,” Journal of Political
Economy, 93: 760–81.
Mehta, Pratap Bhanu. 2002. “Debating Election Finance,” The Hindu (July 17).
Mironov, Maxim and Ekaterina Zhuravskaya. 2014. “Corruption in Procurement and
the Political Cycle in Tunneling: Evidence from Financial Transactions Data,”
Mimeo, Paris School of Economics.
Morlino, Leonardo. 2005. “Anchors and Democratic Change,” Comparative Political
Studies, 38(7): 743–70.
Muralidharan, Karthik and Venkatesh Sundaraman. 2011. “Teacher Performance
Pay: Experimental Evidence from India,” Journal of Political Economy,
119 (1): 39–77.
————. 2013. “Contract Teachers: Experimental Evidence from India.” National
Bureau of Economic Research Working Paper No.19440. Cambridge, MA, USA:
Economic Research (NBER).
Muralidharan, Karthik. 2013. “Priorities for Primary Education Policy in India’s
12th Five Year Plan,” in Barry Bosworth, Arvind Panagariya and Shekhar Shah,
eds, India Policy Forum 2013–2014. New Delhi: SAGE Pulications.
Muralidharan, Karthik, Paul Niehaus, and Sandip Sukhtankar. 2014a. “Building State
Capacity: Evidence from Biometric Smartcards in India,” National Bureau of
Economic Research Working Paper No. 19999. Cambridge, MA, USA: National
Bureau of Economic Research (NBER).
Muralidharan, Karthik, Jishnu Das, Alaka Holla, and Aakash Mohpal. 2014b. “The
Fiscal Cost of Weak Governance: Evidence from Teacher Absence in India,”
Working Paper 20299. Cambridge, MA, USA: National Bureau of Economic
Research (NBER).
Murphy, Kevin M and Andrei Shleifer, and Robert Vishny. 1991. “The Allocation
of Talent: Implications for Growth,” The Quarterly Journal of Economics,
106(2): 503–30.
Niehaus, Paul, Antonia Attanassova, Mariane Bertrand, and Sendhil Mullainathan.
2013. “Targeting with Agents,” American Economic Journal: Economic Policy,
5(1): 206–38.
Niehaus, Paul and Sandip Sukhtankar. 2013a. “Corruption Dynamics: the Golden
Goose Effect,” American Economic Journal: Economic Policy, 5(4): 230–69.
———. 2013b. “The Marginal Rate of Corruption in Public Programs: Evidence
from India,” Journal of Public Economics, 104: 52–64.
Olken, Benjamin A. 2006. “Corruption and the Costs of Redistribution: Micro
Evidence from Indonesia,” Journal of Public Economics, 90(4–5): 853–70.
———. 2009. “Corruption Perceptions vs. Corruption Reality,” Journal of Public
Economics, 93(7–8): 950–64.
Olken, Benjamin A. and Rohini Pande. 2012. “Corruption in Developing Countries,”
Annual Review of Economics, 4(1): 479–509.
PwC. 2014. “Economic Crime: A Threat to Business Globally,” PwC 2014 Global
Economic Crime Survey.
Pande, Rohini. 2007. “Understanding Political Corruption in Low Income Countries,”
Handbook of Development Economics, 4: 3155–84.
Patnatik, Ila and Ajay Shah. 2014. “Reforming India’s Financial System,” Carnegie
Paper, Carnegie Endowment for International Peace.
Peisakhin, Leonid and Paul Pinto. 2010), “Is Transparency an Effective Anti-
Corruption Strategy? Evidence from a Field Experiment in India,” Regulation
and Governance, 4(3): 261–80.
Peisakhin, Leonid. 2012. Transparency and Corruption: Evidence from India, Journal
of Law and Economics, 55(1): 129–149.
Pritchett, Lant. 2009. “A Review of Edward Luce’s In Spite of the Gods: The Strange
Rise of Modern India,” Journal of Economic Literature, 47(3): 771–80.
Programme Evaluation Organization. 2005. “Performance Evaluation of Targeted
Public Distribution System,” Technical Report, Planning Commission,
Government of India.
PRS Legislative Research. 2011. “Legislative Brief: The Public Interest Disclosure
and Protection to Persons Making the Disclosures Bill, 2010,” January 24.
———. 2012. “The Right of Citizens for Time Bound Delivery of Goods and
Services and Redressal of their Grievances Bill, 2011,” September 27.
———. 2013. “Legislative Brief: The Public Procurement Bill, 2012,” September 27.
———. 2014. “Bill Summary: The Prevention of Corruption (Amendment) Bill,
2013,”February 12.
Raha, Swagata. 2012. “State Legislation on Right to Time-Bound Delivery of

Service: An Overview,” Policy Brief, Accountability Initiative, Centre for Policy
Research.
Rajan, Raghuram. 2012. “The Next Generation of Reforms in India,” Speech at
a Function to Re-release a Book of Essays in Honor of Dr. Manmohan Singh,
New Delhi, April 14.
Ravallion, Martin, Dominique van de Walle, Puja Dutta, and Rinku Murgai. 2013.
“Try Telling People their Rights? On Making India’s Largest Antipoverty
Program work in India’s Poorest State,” Mimeo, World Bank.
Roberts, Alasdair. 2010. “A Great and Revolutionary Law? The First Four Years of
India’s Right to Information Act,” Public Administration Review, 70(6): 925–33.
Rodrik, Dani and Arvind Subramanian. 2005. “From ‘Hindu Growth’ to Productivity
Surge: The Mystery of the Indian Growth Transition,” IMF Staff Papers, 52(10):
193–228.
Sastry, Trilochan. 2014. “Towards Decriminalisation of Elections and Politics,”
Economic and Political Weekly, 49(1): 34–41.
Shleifer, Andrei and Robert W. Vishny. 1993. “Corruption,” Quarterly Journal of
Economics, 108(3): 599–617.
Shrinivasan, Rukmini. 2015. “CIC Unable to Make Political Parties Comply with
RTI,” The Hindu, March 17. Available online at http://www.thehindu.com/news/
national/cic-unable-to-make-political-parties-comply-with-rti/article7002835.ece
(accessed on March 18, 2015).
Sircar, Neelanjan and Milan Vaishnav. 2015. “Ignorant Voters or Criminal
Representatives? Why Voters Support Criminal Politicians in India,” Mimeo,
Carnegie Endowment for International Peace.
Siwach, Sukhbir. 2015. “Ashok Khemka, Whistleblower IAS Officer, Transferred
Again,” Times of India, April 2. Available online at http://timesofindia.india-
times.com/india/Ashok-Khemka-whistleblower-IAS-officer-transferred-again/
articleshow/46777082.cms (accessed on April 4, 2015).
Sridharan, E. 2006. “Electoral Finance Reform: The Relevance of International
Experience,” in Vikram Chand, ed., Reinventing Public Service Delivery in India.
New Delhi: SAGE Publications.
Sridharan, E. and Milan Vaishnav. Forthcoming. “The Evolution of India’s Political
Finance Regime,” in Pippa Norris and Andrea Abel van Es, eds, Checkbook
Elections? Political Finance in Comparative Perspective.
Srivastava, Mihir. 2013. “The Congress Bureau of Investigation,” Open, April 6,
Available online at http://www.openthemagazine.com/article/nation/the-
congress-bureau-of-investigation (accessed on March 1, 2014).
Sukhtankar, Sandip. 2012. “Sweetening the Deal? Political Connections and Sugar
Mills in India,” American Economic Journal: Applied Economics, 4(3): 43–63.
Sukhtankar, Sandip. 2015. “The Impact of Corruption on Consumer Markets:
Evidence from the Allocation of 2G Wireless Spectrum in India,” Journal of
Law and Economics, 58(2).
The Economic Times. 2015. “Clean Up Coal, Scrap Flawed Auctions,” March 20.
Available online at http://blogs.economictimes.indiatimes.com/et-editorials/
clean-up-coal-scrap-flawed-auctions/ (accessed on April 2, 2015).
The Economist. 2014. “Campaign Finance in India: Black Money Power,” Banyan
weblog (May 14).
The Hindu. 2013. “Dangerous for Democracy,” March 21. Available online at http://
www.thehindu.com/opinion/editorial/dangerous-for-democracy/article4530365.
ece (accessed on April 2, 2015).
The Times of India. 2015. “All-India Judicial Service Unlikely Soon,” March 15.
Available online at http://timesofindia.indiatimes.com/india/All-India-Judicial-
Service-unlikely-soon/articleshow/46571916.cms (accessed on April 2, 2015).
Transparency International and Centre for Media Studies. 2005. India Corruption
Study 2005. New Delhi: Transparency International India.
Transparency International. 2011. Daily Lives and Corruption: Public Opinion in
South Asia, South Asia Corruption Barometer.
United Nations Office on Drugs and Crime. 2012. India: Probity in Public
Procurement. New Delhi: UNODC.
Vaishnav, Milan. 2012. “The Merits of Money and “Muscle”: Essays on Criminality,
Elections and Democracy in India,” PhD dissertation, Department of Political
Science, Columbia University.
Vaishnav, Milan and Reedy Swanson. 2015. “Does Good Economics Make for Good
Politics? Evidence from Indian States,” India Review, 14(3).
Wade, Robert. 1985. “The Market for Public Office: Why the Indian State is Not
Better at Development,” World Development, 13(4): 467–97.
Yadav, Anumeha. 2014. “Modi Rolls Out Labour Reforms,” The Hindu, October
17, Available online at http://www.thehindu.com/news/national/narendra-modi-
launches-host-of-schemes-at-labour-meet/article6506570.ece (accessed on
April 6, 2015).
Comments and Discussion
Devesh Kapur
University of Pennsylvania
This is a carefully argued and impressively documented paper on a complex

and contentious subject. Any analytical work in this area faces severe data
and measurement challenges given the illegality of the subject on the one
hand and the data requirements to establish magnitudes, effects, and precise
causal mechanisms on the other hand. Yet given the importance of the issue,
the authors’ are to be commended for taking on this difficult challenge in a
lucid and comprehensive manner. As is the case with all good papers they
help stimulate further thinking and my comments are in that spirit.
The paper defines corruption as “the misuse of public office for private
gain” (which it attributes to Pranab Bardhan). A sharp definition is essential
not just for analytical clarity but also for tractability given the vast terrain
the issue entails. But what do the two core components of this definition—
“public office” and “private gain”—imply for understanding the nature of the
problem? Take the example of two faculty, one of who teaches at a private
university and the other at a public university. Now, if both give grades in
exchange for money, by definition only the faculty in the public university
is corrupt because he is a public official but the one in the private university
is not. The latter might be guilty of something else but not corruption even
though the actions are identical.
By defining corruption in a particular way it means that our understanding
of corruption-like activities in the large—and growing—terrain of economic
activity in the private sector is severely constrained. In the example cited
above—higher education—the private sector is the dominant actor and its
operations are rife with fraudulent activity. In the case of “grand corrup-
tion” in particular the private sector is integral to public corruption. Two
decades ago when a minister took a bribe of `2 crores, he could hide it in
suitcases. But when bribes run to many multiples of this amount it cannot
be hid in suitcases. It must be laundered and the mechanism of laundering
is the private sector.
Widespread malfeasance in the private sector often ricochets back to the
state. The sordid affairs of statutory regulatory bodies of self-governance
such as the Medical Council of India or the Bar Council of India, none of
262
which are state run institutions, have had severely pernicious consequences
for the medical and legal professions. This is also true of many NGOs which
are often seen as heroically battling against the depredations of the state.
Many indeed play that role but equally many non-profits are conduits for
money laundering.
The other part of the definition which also needs a deeper understanding
is the idea of “private gain.” What exactly is private gain? Is private gain
equivalent to pecuniary gain? So let us take a hypothetical country with a
hypothetical prime minister who is completely honest but decides to look the
other way when her or his ministers are engaged in patently corrupt activ-
ities in order to hold on to power. Is that an act of commission or an act of
omission? If we care about welfare should we worry more about someone
who takes 50 crores in illegal money or someone who has the power to stop
such activities and allow certain policies that cost 500 crores to the public
exchequer? If the ambition in public organizations to stay in power can
sometimes cause a lot more damage than someone taking cash, then it begs
the question of what exactly is “private gain” and what types of “private
gain” should we focus on?
Needless to say expanding the definition of corruption would mean an
entirely new paper. But it would be good to at least examine the implica-
tions of defining corruption in a very specific way for our understanding of
this complex issue.
The main thrust of the paper is addressing four big questions:
1. The underlying causes of corruption with the authors’ identifying

the mismatch between regulatory complexity and weak enforcement
capacity as the root institutional factor shaping India’s corruption
environment and election finance and weaknesses in public sector
recruitment and transfers as more “proximate” causes.
2. Different types of corruption—facilitative, collusive, and extractive—
and the magnitudes, causes, and consequences of each of these types
of corruption.
3. Strategies for combating corruption, ranging from introducing new
technologies to legal changes.
4. India’s political economy of reform which might shape how India
addresses this scourge.
On the first question—the severe mismatch between regulatory complex-

ity and enforcement capacity—there are two axes of variation that future
work could exploit: variation across the principal domains of activity of
a modern state and variation on each of these domains across states and
between the center and states. There are four key domains of activity of the
modern state: the public goods and welfare state where leakages have been
galore and have been the subject of numerous studies ranging from PDS
to NREGS, etc…. The second is the regulatory state which has been the
source of “terrestrial rents (from the allocation of land), subterranean rents
(from the allocation of rights to coal mining and oil and gas exploration)
and ethereal rents (from the allocation of spectrum)”(Subramanian 2012).
The consequences have been both distributional (with the consumer pick-
ing up the costs) as well as efficiency costs, with severe adverse effects on
the supply capacity of the economy. The third domain of state activity is
the extractive state at the heart of which lies taxation. Although there is a
widespread belief that there is large-scale rent seeking in this domain, it is
not addressed in the paper.
The fourth (which is addressed in the paper), is the law and order state,
in particular the police and courts. The paper emphasizes how the weak-
ness in enforcement has hampered India’s ability to address corruption. If
addressing corruption requires competent investigative and prosecutorial
machinery, asking the very group that is amongst the most corrupt parts of
the Indian state—the police—to address corruption, is unlikely to bear fruit.
Quis custodiet ipsos custodies—who guards the guardians?—is a conundrum
intrinsic to all power, but in India the dilemma has been made more acute
by the virtual absence of serious reforms in the law and order institutions
of the state (especially the police) since independence. While the paper
draws attention to this reality, it raises a deeper question on why this issue
has remained unaddressed for so long. The politician–bureaucrat–police
nexus is the usual answer, but it raises the question why in such a highly
competitive democracy as India’s, police reform has not had much issue
salience among voters.
The paper draws attention to the distributional and efficiency costs of cor-
ruption. But it is noteworthy that when the major corruption scandals under
UPA-II came under scrutiny—by the media, the courts, and the Comptroller
and Auditor General—growth declined as policy paralysis in the govern-
ment brought decision making in the government to a halt. Indeed the costs
stemming from the ensuing decline in growth rates far exceeded the costs of
all the corruption scandals. Obviously this does not mean that the two were
directly causal or even if they were that corruption is a price worth paying
for higher growth rates given the likelihood of severe long-term institutional
costs. The paper points to the reality of a certain type of democratic account-
ability in the Indian system. When major corruption scandals erupt in the
public view, the decisions are often reversed. But the resulting slowdown
in decision making not only significantly raises costs (especially in capital
intensive projects) but has troubling non-pecuniary costs such as the effects
of the paralysis in defence procurement on India’s defence preparedness.
This points to an issue that needs better understanding in the Indian con-
text: the difference between “efficient” and “inefficient” corruption, often
seen as the difference between corruption in East Asia and Africa, or in the
Indian case between a Tamil Nadu and Uttar Pradesh. It is not that corruption
is less in Tamil Nadu but things actually get done there unlike Uttar Pradesh.
Corruption in public contracts in India when coupled with a rigid adherence
to L1 (lowest) bidding in procurement results in low-quality public infra-
structure. In East Asia corruption results in higher procurement costs while
maintaining quality and the resulting longer life-span of the product means
that long-term costs are much less. In recent years (as the paper mentions)
there are reports of politicians taking equity stakes in projects (instead of
upfront cash payments). This is a form of credible contracting that aligns
the incentives of the politician with that of the project promoter and ensures
better project outcomes compared to cash bribes. Given the strong resist-
ance of politicians to curbing corruption, even moving from “inefficient” to
“efficient” corruption is likely to bring substantial efficiency gains.
The paper relies heavily on a wide range of survey experiments to come to
broad conclusions. They are for the most part ambiguous not just for external
validity reasons but even their internal logic. It is not just that most of these
experiments are over very small samples. Even where they are impressively
large, the policy implications are ambiguous. Take the case of salary incen-
tives. Higher salaries in principle appear to have a positive selection effect
on who joins public service (on measures such cognitive skills, personality
traits, and motivation) ( Ernesto Dal Bó and Rossi 2013). Yet, what they do
after they join is a function of host of other institutional and organizational
factors. If public teachers in India who are already paid much more than
their private school counterparts are offered further financial incentives for
performance and their performance improves (as in the case of teachers in
AP), it is unclear if this is a new equilibrium or will it slide back as time
passes and researchers (and scrutiny) have moved on and whether other
parts of the public sector will demand similar deals without commensurate
long-term improvements in performance. In other cases experiments that try
and gauge the impact of information on say voter behavior simply cannot
distinguish between the information given by the surveyor and the actual
understanding of that information by the respondent.
It is an open question whether the analytical methods relative to the

issue being addressed are a bit like a garden hose for a large forest fire, and
how meaningful they are to address an issue as deeply systemic and which
requires systemic policy options. Researchers tend to develop partisan stakes
in their results which often closes them to thinking about changes over time.
To take an example, amidst the many contentious debates on leakages in the
PDS, several papers argued that there had been a turn around, singling out
the case of Chhattisgarh which emerged as the poster state for the public
distribution system. And yet in early 2015 stories emerged of vast corruption
in the Chhattisgarh State Civil Supply Corporation (the nodal authority for
the public distribution system in the state) with an estimated 14 lakh fake
ration cards issues since the enactment of the state’s much-acclaimed Food
Security Act in January 2013 (Indian Express 2015).
The paper focuses on six broad categories of tools to combat corrup-
tion: (a) Information/ bottom-up monitoring, (b) Technology, (c) Financial
incentives/ performance pay/efficiency wages, (d) Electoral reform/political
incentives, (e) Legal reform, and (f) Policy reform. The authors’ are rightly
skeptical whether information and community participation on its own will
make a dent on addressing corruption and advocate putting more bargaining
power in the hands of beneficiaries through measures such as the passage
of “Right to Service” bills. But by focusing on information provided by
researchers to respondents the paper misses out on the big picture, namely
the role of the media. While the media’s coverage of corruption “scams”
may indeed be “sensationalist” (as the paper argues), the counterfactual—
whether the government would have acted in the absence of the media’s
watchdog role—is very unlikely. Indeed the role of the electronic media—
and increasingly social media—in shaping political responses to ensuring
some accountability and enforcement is a rich research area.
Yet all of these might matter only so much in the absence of far greater
penalties on erring officials. There is a need for better research on the struc-
ture of penalties in public service, from fines to career paths to dismissal,
and the institutional changes required to make them more onerous. But rede-
signing public institutions is more than simply a matter of incentives such
as salaries and penalties. It is about designing better systems of screening
and selection, training and work culture, improving organizational cohesion,
and autonomy from political pressure. Here again, ethnographic case stud-
ies of specific organizations are needed to better understand organizational
behavior.
Finally there is a big puzzle. Everyone knows that corruption in India is
ubiquitous and recognizes that it is a serious problem. So why isn’t much
done about it? Is it because corruption is an inevitable penalty that India

must pay to accommodate and sustain its democratic system? On the one
hand this generates ever growing needs for election financing, which like
a gigantic tornado sucks in enormous amounts of resources which have to
be generated from corruption. On the other it might simply be that a rigidly
hierarchical society in the process of complex changes provides the struc-
tural conditions for the tornado to sustain itself. Coalition governments are
seen to be more susceptible to corruption. But coalition governments are
a reflection of political fragmentation, which in turn is (at least to some
degree) related to India’s social heterogeneities. Hence, if the roots of the
problem lie not so much in politics or business but in Indian society itself,
the solutions might lie in the nature and pace of social change.
Karthik Muralidharan
University of California, San Diego and NCAER
It is a pleasure to be here. I think that this is a very nice paper and as Devesh
said there is a very vast and disconnected body of evidence that it makes
a valiant attempt at bringing together, and it does a good job putting these
things in one place and I think some sections are particularly strong. For
example, I particularly like the discussion on the effectiveness of provid-
ing information, and the discussion of why this may not always be enough.
There is also this nice additional section that creates a list of major scams
and organizes by sector just to help better understand the areas that are most
vulnerable to corruption. But, I think it may work better in the beginning to
motivate the rest of your paper.
I think the paper has the potential to be a go-to reference for those want-
ing to get up to speed on the rapidly growing literature for both academic
and policy audiences but I think it needs a fair bit of work. The good news
is that most of this is organizational and conceptual and I am at least not
going to ask or expect you to go collect more primary data which is kind
of not realistic but I think based on what you have you can make the paper
a lot more punchy by doing some more organizational stuff. I am going to
have three classes of comments.
The first is on organization and framework. The second is on topics
covered and excluded and the third is on interpretation of some of the
results that I think right now reads a little bit too much like a laundry list:
putting them together in a framework that helps make sense in a more first
principles kind of way.
There is also one thing I didn’t mention here because Devesh did. So let
me just mention that this is important and is almost worth having a little
section on what I would call forensics. So the fact that corruption is really,
really hard to measure means that there is a class of papers that is important
just because they manage to measure convincingly the fact that corruption
takes place. Ray Fisman’s Indonesia Political Connections paper (Fisman
2001) is probably the best example of this, but even your sugarcane paper
(Sukhtankar 2012), and Devesh and Milan’s cement paper (Kapur and
Vaishnav 2015) are important papers even though they are not like super
well identified, just because they get at this very difficult task of measure-
ment. So I think calling this section “forensics” is a useful way to do this
because it is really about measurement and detective work as much as it is
about theories and causality.
But coming to the economics of this I think the one really important
paper that you are missing in terms of helping organize your thoughts is the
Bandiera et al. (2009) AER paper on active and passive waste in govern-
ment. The main insight of this paper is that while most of the attention goes
to big headline grabbing scams—they estimate that about 80 percent of the
waste in government is because of what they call “passive waste” which is
not the kind of corruption that is directly going into somebody’s pockets
but the waste that happens because of systemic inefficiencies, and of the
waste that happens because there is no residual claimant in government to
internalize the returns to improving effectiveness of governance and func-
tioning. Normally, I would say maybe this is too much and this might be
biting off more than you can chew (and that you should stick to corruption
defined as “active” waste above).
But because you have chosen to discuss both scams as well as the day-
to-day (or annual) losses, you almost certainly want this framework because
the policy implications are quite different. So, I think having that would be
useful.
There is also a second distinction that is similar but subtly different.
This is the difference between a corrupt state and a nonresponsive state,
and so it is related active and passive waste and they are correlated but you
can see why there is a different policy implication. For example, I think
when you think about the RTI/RTS, the problem that these acts are trying
to solve at some deep level is not necessarily corruption as much as just a
nonresponsive state. It is true they might be correlated because there might
be deliberate denial of service to create the bottlenecks to create the rents
in the first place. But the policy implications are different: For example,
once you put in place a “Right to Service” act and monitor progress with
electronic tracking of files, one of the things it helps identify is cases of
genuine staffing shortages in the system or cases that can be improved by
simple operations improvements.
So for instance, the former collector of Hyderabad would talk about a very
simple example about how students would want income and caste certificates
at the time of college applications, and they know that 10 months of the year
there is no demand and two months of the year there is this absolute peak
demand. So even if nobody is corrupt the service is going to be very slow
at the point of peak demand, which is the time when the typical beneficiary
interacts with the government for this service. So the nonresponsive state
might just be reflecting understaffing or mismatch of average and peak
demand/supply, which can be improved a lot with just simple operational
efficiencies. If you look at police, e.g., they are chronically understaffed and
under-resourced, and I will come back to the connections between corrup-
tion because there is this issue of deliberate under-investment in state capa-
city which I think is really important point. But I think your discussion of
RTS/RTI will really benefit from this distinction.
A final point, which is not made, is distinguishing between the class of
studies and recommendations that you would make to a motivated policy-
maker who says: “I want to reduce corruption” versus discussing the political
incentive constraints for reducing corruption in the first place. This impor-
tant because the binding constraint is often not whether policy X works but
whether policy X was allowed to be implemented in the first place. So we
understand this and your review again has elements of each of these. But if
you make the distinction clear and highlight the connections where relevant
I think it will make the whole thing much more coherent and I will come
back with concrete examples of this.
So choice of topics I think Devesh has talked about this. Some sections
really do not belong in this review. Section 2.4 we have talked about this. It
is like seven pages and it is really like calling democracy corruption, means
like it was thrown in there because there is evidence in that place but I don’t
think it belongs at all. The one place I think where it might belong is if the
motivation of that section is to use the existence of tainted politicians as a
reason for why anticorruption policies might never get implemented. I think
it was Chitra Subramaniam in the context of Bofors who once famously said
every politician loves corruption scandals only to the extent that it helps
show the other guy in a bad light but nobody is actually interested ever in
in the first place. But the policy implications are different: For example,
once you put in place a “Right to Service” act and monitor progress with
electronic tracking of files, one of the things it helps identify is cases of
genuine staffing shortages in the system or cases that can be improved by
simple operations improvements.
So for instance, the former collector of Hyderabad would talk about a very
simple example about how students would want income and caste certificates
at the time of college applications, and they know that 10 months of the year
there is no demand and two months of the year there is this absolute peak
demand. So even if nobody is corrupt the service is going to be very slow
at the point of peak demand, which is the time when the typical beneficiary
interacts with the government for this service. So the nonresponsive state
might just be reflecting understaffing or mismatch of average and peak
demand/supply, which can be improved a lot with just simple operational
efficiencies. If you look at police, e.g., they are chronically understaffed and
under-resourced, and I will come back to the connections between corrup-
tion because there is this issue of deliberate under-investment in state capa-
city which I think is really important point. But I think your discussion of
RTS/RTI will really benefit from this distinction.
A final point, which is not made, is distinguishing between the class of
studies and recommendations that you would make to a motivated policy-
maker who says: “I want to reduce corruption” versus discussing the political
incentive constraints for reducing corruption in the first place. This impor-
tant because the binding constraint is often not whether policy X works but
whether policy X was allowed to be implemented in the first place. So we
understand this and your review again has elements of each of these. But if
you make the distinction clear and highlight the connections where relevant
I think it will make the whole thing much more coherent and I will come
back with concrete examples of this.
So choice of topics I think Devesh has talked about this. Some sections
really do not belong in this review. Section 2.4 we have talked about this. It
is like seven pages and it is really like calling democracy corruption, means
like it was thrown in there because there is evidence in that place but I don’t
think it belongs at all. The one place I think where it might belong is if the
motivation of that section is to use the existence of tainted politicians as a
reason for why anticorruption policies might never get implemented. I think
it was Chitra Subramaniam in the context of Bofors who once famously said
every politician loves corruption scandals only to the extent that it helps
show the other guy in a bad light but nobody is actually interested ever in
doing anything about it and I think that is the one area where those two get
connected. But otherwise it is mostly a different literature.
The other important topic is not covered and to me the single biggest
omission is corruption in public sector recruitment and I think it is important
both at the highest levels, and at the entry level, and not just in terms of
recruitment scams which are not in your taxonomy. The recruitment scams
are not in your list and we now have several prominent examples includ-
ing the Madhya Pradesh scam, or fake certificates for teacher recruitment
in Bihar, there is the teacher recruitment scam in Haryana, where a former
Chief Minister (Om Prakash Chautala) actually got convicted, and not to
mention the recent resignation of the Railway minister on similar charges.
So we have evidence of these scams but actually theoretically this might
in fact be the most pernicious form of corruption, more so than individual
scams and that is because you can write a very simple overlapping genera-
tion model to show how once you get a bad egg into the system in a senior
place that this is going to perpetuate into eternity because the guy who has
bought his way into the job has to recovery money and so it is much worse
than an individual kind of corruption because you get negative selection
and so if you want to think about what is that causes institutions to decay
at some level, it fundamentally, in my view, comes down to corruption in
public sector recruitment.
So you get negative selection in terms of who comes in, and you change
the norms, because if you paid for your job, serving the public interest is
going to be lower on your priority list than recovering your “investment”
through corruption. Sometimes Hindi movies have a way of capturing
insight better than most academic seminars. How many people have seen
the movie Gangajal? There is a great scene, where you have got this corrupt
cop, holding up highway traffic and extorting money and then when this
honest SP catches him red handed and tries to take action he comes back
and cries “Saheb I have paid `3 lakh for this job.” I have taken a loan for
this job and I have no other option. It just kind of in a very visceral way
shows how the public sector recruitment, in my view at least, is probably
the foundation of almost all of these problems in linking corruption and poor
service delivery. I think there is a different set of issues at the top which is
the discretion that officers have at the top. So when people are willing to
pay to become DGPs and members of railway board, that is a different kind
of corruption than for jobs at entry level where I think the real problem is
just that the wages are too high. So we have got a wage compression in the
public sector where over 95 percent of public sector employees are paid
massive multiples of what the open market wages are but the very senior
public sector employees (in terms of skills) are massively underpaid relative
to their private sector opportunity cost and this leads to distortions at both
ends of the salary distribution.
Here is just one presentation to illustrate this. This is from my work on
contract teachers where the education establishment thinks that these con-
tract teachers (who are hired in the village —young women, less educated,
less trained, paid 1/6 of a civil service teacher is paid) are all being exploited
relative to government teachers. But if you then look at the private sector
wage, if anything the private salaries are even lower. So it is not like contract
teachers are being exploited but that there are massive rents in the civil ser-
vice job. In the past, I used to just think this was fiscal inefficiency because
what my work shows is that contract teachers are equally effective as regular
teachers paid many times more. If anything, they appear even more effective
because they do not have civil service tenure and they are connected to the
village. In the past, I just used to feel irritated that this is a fiscal waste but I
increasingly think it is worse than waste and that is because of this negative
selection problem. You can never really hide from the market. The market
will come back and bite you even though you don’t expect it. So when you
have this kind of wage premium what is going to show up is there is market
for getting these jobs and so that is what all these teacher recruitment scams
are about and people are willing to pay these huge premia but you just are
capitalizing upfront, a lifetime of rents that is above the market clearing
wage and so you cannot really hide from the market and that is why I think
this issue of wage compression is so important.
The second key topic which I have already alluded to is this: The great
value of adding a political scientist to a paper on corruption is the possibility
of a more nuanced understanding of the political conditions under which
anticorruption reforms actually can take pace and I will come back to that.
My last presentation: some connections and interpretations which I think
are really useful to make. The first connection is between technology and
incentives. I think it is important to make connection because I think technol-
ogy makes it easier to measure and implement incentives but you still need
the incentives. The most illustrative case of this is the empirical journey from
first draft to publication of the famous “Cameras paper” (Duflo, Hanna, and
Ryan 2011). So everybody thinks about this as the “Cameras paper,” that
you put cameras in school and teacher attendance went up. Actually that is
wrong. So if you look at the working paper version of this paper it had the
title Monitoring Works in 2005. By the time it got published in the year 2012
the title had changed to Incentives Work and the subtle difference is that
the intervention combined a camera and a wage schedule that was paying
you every day for valid days in which you showed up and so eventually
what happens is they can use certain structural features of the compensation
formula to isolate whether the impact is from the monitoring or from the
fact that you are actually paid for each day you show up and once they do
that the action shifts completely from the monitoring, to the incentives itself.
This is not to say that technology does not matter but the important point
the technology is doing from a contract theory perspective is it is making
the effort not just observable but verifiable. So the difference between an
individual inspector who goes to the school and finds the teacher is not there
but then has to prove that in a court of law (at which point it is his word
against the teacher’s word) is that the camera makes it verifiable to a third
party that the teacher was absent, which reduces the cost of taking the fol-
low up action. So technology is a really important facilitator for reducing
the cost of implementing incentives, sort of like George Baker’s work on
efficiency in the trucking industry. Once you are able to track the move-
ment of trucks that reduces the rents that truck drivers were getting from
long periods of slack because now you can track exactly where the truck is
going. So the technology is an enabler in terms of reducing the measurement
and monitoring cost but you need the incentives on top of that to improve
service delivery.
So the other important point that has come up in this discussion this
difference I think between NGO implementation versus government imple-
mentation of pilots. So we have now got multiple studies by Abhijit Banerjee
and co-authors, and a similar study in Kenya, and a common thread is that
when you have pilot projects implemented by researchers and non-profits
they seem to work. But then when you try to transfer this to government
implementation it kind of breaks down. I think it is a little more nuanced
than that because this is in contrast to kind of two sets of studies we have.
One is the example of our recent work with the Government of AP where
they randomized the deployment of biometric Smartcards to make NREGS
payments across 20 million people. This was implemented by the govern-
ment and we find big positive effects on reducing corruption. Second is our
recent work on teacher absence which we just released yesterday and we
are not touching any program. We are just looking at the impact of good
old monitoring by the government machinery (measured by whether your
school was visited in the past three months) and we find that that has a big
correlation with reduced teacher absence. So what I think is going is very
simple. The difference in each of these cases is that in the earlier cases, the
impetus for the change came from outside the government from researchers
and maybe one motivated IAS officer and then once that guy got transferred
there was no institutional desire to implement the change whereas in these

two cases both of these studies are evidence from cases where the govern-
ment had made a policy decision on its own to implement a certain kind of
improved implementation capacity intervention and there we find systematic
effect. So the good news is that when you do these things in a way that the
“system has bought in” it does seem to work and so that begs the question
of what it takes for the “system” to buy in to implement an anticorruption
initiative. To be fair, I think the smaller NGO studies have gotten a bad rap
but the way to think about them is very much like in medicine. We have
efficacy trials and effectiveness trials. So the efficacy trial is testing whether
something works under good implementation, and then the effectiveness
trial is when you take that to implementation under “typical conditions” and
see whether it works. So the smaller scale studies give us the first step, but
then you need systems to adopt the findings and embed them into practice.
The last point is what this really does is it begs a different question about
what are the kind of political incentives for investment in anticorruption
and I think the last piece that I would really think the paper would benefit
from is a discussion of the political incentives deliberate under-investment
in state implementation capacity and the fact that this creates scarcity which
then makes the politician the arbitrator of who in fact gets access to scarce
public resources. I think this is a key connection between political corrup-
tion and service delivery. Some of the best work on this has been done by
former IAS officers who have then gotten PhD’s. So whether it is Santosh
Mathew or Anirudh Krishna they have done work on deliberate under-
investment by the state. Anirudh had this wonderful terminology to describe
political incentives to develop what he calls “pooch” and “pahunch,” which
roughly translate into “recognition” and “influence” respectively. So, what
a politician cares about is “who is coming to me for favors?” and that is
his poonch and the pahunch” determines the extent to which he can deliver
on that request. So the response of the politician to ideas that may improve
service delivery (with say new technologies) is to say: “Meri poonch kahan
rahegi?” that translates into worrying that “nobody is going to come to me
for anything if the system is going to deliver these services”.
I think this is a very useful point to keep in mind, and it raises the question
about the conditions under which states will choose to invest in improved
capacity for lower corruption and better service delivery, and to ask what
can be learnt from the historical evidence of how societies move from a
politics of patronage to a politics of service delivery. Then the research we
have on interventions and their effectiveness work that might feed us inputs
into politicians who want to make that change.
General Discussion
Pranab Bardhan started by noting that the paper could be tighter on defini-
tions of corruption, even though they would inevitably overlap. The defini-
tion he himself used focusing on public corruption only was useful because
it allowed him to confine himself to just one kind of corruption. He then
referred to two other aspects he felt are important: the impact of competition
from private providers contracted to provide public services (e.g., passports)
on reducing corruption; and, second, wholesale versus retail corruption (e.g.,
some East Asian countries versus India) where the former is more “efficient”
and does not affect the decision at the margin, but of course leads to political
and institutional problems.
Dilip Mookherjee noted that the some societies deal with corruption by
institutionalizing it, e.g., US campaign finance laws that require disclosure
but legally allow contributions in an institutionalized form that many might
consider corruption. On looking for remedies, he asked if they were studies
in India that looked at the impact of a scam on the political career of the
politician involved even if they were not convicted. He also wondered
whether greater decentralization would be a solution, citing his own work
in West Bengal with Pranab Bardhan. He noted that reducing the political
discretion enjoyed by bureaucrats and politicians would be good for reduc-
ing corruption, citing the example of the USA. Finally, he noted that the
evidence is still out on whether community monitoring and provision of
information could curtail corruption.
Sanjeev Ahluwalia noted that while the paper had a lot of useful insights
on corruption he felt that many of them were too generic. He felt the paper
would gain from looking at measures that strengthen the social compact
between the state and citizens, e.g., broad basing taxes and increasing usage
charges. These sorts of measure would allow people to understand better
that corruption is simply the theft of their own money.
Amartya Lahiri said that corruption in India has possibly become so
endemic to society that it is simply accepted as a way of doing business.
Narrow, programmatic interventions working on a specific dimension of
corruption might therefore not work well since the problem may be so syste-
matic. Simple interventions would likely lead to multiple equilibria, with
spillovers from the public to the private sectors and vice versa.
Ashok Lahiri said that we could and should look at the Singapore model,
where bureaucrat salaries are high, but would raising salaries alone solve the
problem in India? What about recruitment and hiring and firing practices?
Nirvikar Singh questioned the notion that corruption was just a transfer
payment and redistributive and therefore may have no resource allocation
costs. He suggested that when there is a misallocation already then corrup-
tion can have serious resource allocation costs and in that situation it would
also have a long run impact on growth, though it is hard to estimate that
impact. He suggested that corruption also eroded trust and degraded social
norms, therefore substantially increasing its cost to society. This is not easy
to estimate but is nonetheless very important.
Mihir Desai seconded Karthik’s comment that the paper had the potential
of being a go-to reference, but, apart from policy recommendations, wanted
to see it end with many larger questions that researchers should address and
on which we don’t really know enough. He also cautioned against viewing
mismanagement as corruption, because otherwise almost everything would
be corrupt. For example, in his view, teacher abstenteeism was mismanage-
ment and not corruption. He disagreed with Devesh Kapur that measurement
of corruption was important in itself and we could stop there, and suggested
that measurement should be complemented by theory if we are to look for
solutions. Finally, while routine day-to-day corruption was more costly than
imagined, he urged the authors not to discount the costs of high-level cor-
ruption because of its systematic effects, which Nirvikar had also referred to.
Arvind Panagariya also emphasized the need for a unifying theme.
He suggesting looking at policy induced corruption that goes away when
policies are changed. He felt that technology offered solutions that could
reduce corruption dramatically, as for example in the case of railway ticket
booking when it was computerized. He suggested that one way of organ-
izing the paper might be to look at the difficulty of enforcement in India,
an important cause of corruption.
Anupam Khanna felt that this was the most informative and useful
discussion of corruption that he had heard in 20 years. He felt that a lot of
corruption in India was now at the government-to-business and the business-
to-business interfaces. On G2B, there is a real need to understand better how
corruption happens and then what to do about it. The popular perception of
the IT sector is that it is very clean, and in many ways it is, but there are pay-
offs there too, and while it is hard to clearly classify this as corruption, there
is something wrong with it, and it is important to understand what and why.
Rajnish Mehra (Chair) said that it would be useful to have a sense of the
magnitude of the seven categories of corruption noted in the paper: what is
first order, what is second order? On strategies to combat corruption, he felt
that it would be good to see how they are matched to corruption categories.
He felt that corruption distorted the investment-consumption tradeoff, which
then affects an economy’s growth rate, and that this impact was far more
important than suggested in the paper. So, one implication of this way of
thinking would be to see which of the seven categories in the paper are likely
to change growth rates and others that are more redistributive in their impact.
Karthik Muralidharan noted the massive wage compression in the pub-
lic sector: relative to the private sector, the top 2 percent of government
employees are massively underpaid and the vast majority of government
employees are paid a massive multiple. The Singapore model is about paying
higher salaries but with immediate exit if you are corrupt, so that it creates
an efficiency wage model. The reason why efficiency wage models do not
work in India is that one never gets fired. So, it is not just the level of pay
but the structure of pay that matters. He pointed to his work in Indonesia
where they doubled teacher pay across the board and there was absolutely
zero impact on learning outcomes, but when there was a tiny performance
pay, just 3 percent of their salary, now linked to performance, then they got
dramatic results.
References
Subramanian, Arvind. 2012. “What is India’s Real Growth Potential?” Business
Standard, May 23.
Ernesto Dal Bó, Frederico Finan, and Martín Rossi. 2013. “Strengthening State
Capabilities: The Role of Financial Incentives in the Call to Public Service,”
Quarterly Journal of Economics, 128(3): 1169–1218.
Indian Express. 2015. “Fake Ration Cards Pile Up in ‘Model’ PDS State,” Indian
Express, April 9. Available online at http://indianexpress.com/article/india/
politics/fake-ration-cards-pile-up-in-model-pds-state/ (accessed on April 30,
2015).

d4

Uploaded by

Copyright:

Available Formats

d4

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

d4

Uploaded by

Copyright:

Available Formats

SANDIP SUKHTANKAR*

Corruption in India: Bridging Research

ABSTRACT Corruption has become an increasingly salient issue in India today,

Keywords: Corruption, India, Transparency, Public Sector, Political Financing,

JEL Classification: D4, D73, H10, H40, H83, K42, O10

C orruption in India is a topic that seems to never fall out of fashion.

officials in order to obtain contracts or licenses from the government, then

such reviews do not produce explicit recommendations for formulating

In this section, we briefly review the major causes of corruption in India.

those countries standing at the unfortunate juncture of a heavily regulated

2.1. Regulatory Complexity

2.2. Lack of Enforcement and Implementation Capacity

FIGURE 1. Public Sector Employment in G-20 Countries

FIGURE 2. Public Employment in India from 1971–2001 (millions)

Compounding the administrative burden they face, officers regularly face

TABLE 2. Police Vacany Rates across Indian States, 2013

The judiciary faces the same dilemma: almost 30 percent of seats on

FIGURE 3. Unfilled Vacancies in Judicial System

2.3. Inadequate Regulation of Political Finance

FIGURE 4. Outstanding Cases in Indian Courts, 2013

forces, politicians and business will have reason to perpetuate a system of

In theory, if parties were transparent in their finances, some triangulation

Sukhtankar’s paper captures an important truth characterizing India’s

34 percent face pending criminal cases while 20 percent face charges of a

2.4. Public Sector Recruitment, Postings, and Transfers

It has long been known that recruitment, transfers, and promotions

3. Varieties of Corruption: Magnitudes and Consequences

Having reviewed some of the leading drivers of corruption, in this section we

and bribes paid to illegitimately obtain government contracts or licenses

3.1. Facilitative Corruption

Theoretically, this type of corruption likely comes under the heading of

75 percent of households reported paying bribes to obtain the cards, although

3.2. Collusive Corruption

Unfortunately, this type of corruption is also difficult to empirically

The motivating premise of their study is that politicians often turn to

3.3. Extractive Collusion

Embezzlement has distortionary consequences for optimal public finance.

in Jaunpur district in Uttar Pradesh, and Banerjee, Duflo, and Glennerster

large fiscal cost to the government: Muralidharan et al. (2014b) estimate

4. Strategies and Legislation to Combat Corruption:

The previous section catalogued the categories of corruption scholars have

1. Information/ bottom up monitoring

4.1. Information/Bottom-up Monitoring

10. http://www.transparency.org/whatwedo/activity/action, accessed June 25, 2014.

dwellers (NREGA), access to basic education (Right to Education), and the

Providing information as well as encouraging community-based monitor-

of education in villages, and trained volunteers in teaching basic reading

private sector.”16 Former UPA Finance Minister P. Chidambaram called it

and password-protected time-stamp machines and asked to stamp cards at

payments through local customer service providers (CSPs) who were

large variation in access to Internet technology or the continued existence

4.3. Financial Incentives/Performance Pay/Efficiency Wages

4.4. Electoral Reform/Political Incentives

To curtail the entry of candidates associated with illegal activity in the

recommended that candidates against whom charges have been framed by

A second remedy under consideration is to enhance the ECI’s authorities

The null result on the corruption measure, however, complicates the

4.5. Legal Reform

4.6. Policy Reform

order to investigate central government officials on state deputation, as is