Asset disclosure and political selection: Evidence from India [VERY PRELIMINARY AND NOT FOR CIRCULATION] * Raymond Fisman † Florian Schulz ‡ Vikrant Vig § This version: September 2015 Abstract We study the effect of financial disclosure on the pool of politicians contesting elections, and their asset accumulation while in office. Our empirical design exploits the stagger- ing of Indian state assembly elections to credibly distinguish the effect of disclosure laws from broader time trends. We document a large (13 percentage point) one-time increase in exit of winning candidates post-disclosure. The resulting selection coincides with in- creased winning probability of remaining incumbents, relative to a set of counterfactual candidates, suggesting that low-ability candidates exit as a result of disclosure. The pri- vate benefits of public office, as measured by relative wealth growth premium of election winners versus runners-up, decreases for more experienced politicians (incumbents and ministers) over the subsequent election cycle. For politicians who remain in office across multiple elections, we also find a decline in asset growth, indicating that disclosure laws af- fected politician behavior in office. Overall, our findings suggest that mandatory financial disclosure resulted in the self-selection of more desirable (by the electorate) candidates, and possibly reduced rent extraction by office-holders. JEL Classification : D72; D73; D78 Keywords : Information disclosure; Political selection; Indian politics; * Acknowledgments: We would like to thank Andrew Siegel for helpful comments and suggestions. † Boston University. Email: [email protected]‡ University of Washington. Email: [email protected]§ London Business School. Email: [email protected]
39
Embed
Asset disclosure and political selection: Evidence from India … · 2020-05-30 · Asset disclosure and political selection: Evidence from India [VERY PRELIMINARY AND NOT FOR CIRCULATION]
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Asset disclosure and political selection: Evidence
from India [VERY PRELIMINARY AND NOT
FOR CIRCULATION]∗
Raymond Fisman†
Florian Schulz‡
Vikrant Vig§
This version: September 2015
Abstract
We study the effect of financial disclosure on the pool of politicians contesting elections,
and their asset accumulation while in office. Our empirical design exploits the stagger-
ing of Indian state assembly elections to credibly distinguish the effect of disclosure laws
from broader time trends. We document a large (13 percentage point) one-time increase
in exit of winning candidates post-disclosure. The resulting selection coincides with in-
creased winning probability of remaining incumbents, relative to a set of counterfactual
candidates, suggesting that low-ability candidates exit as a result of disclosure. The pri-
vate benefits of public office, as measured by relative wealth growth premium of election
winners versus runners-up, decreases for more experienced politicians (incumbents and
ministers) over the subsequent election cycle. For politicians who remain in office across
multiple elections, we also find a decline in asset growth, indicating that disclosure laws af-
fected politician behavior in office. Overall, our findings suggest that mandatory financial
disclosure resulted in the self-selection of more desirable (by the electorate) candidates,
and possibly reduced rent extraction by office-holders.
JEL Classification: D72; D73; D78
Keywords: Information disclosure; Political selection; Indian politics;
∗Acknowledgments: We would like to thank Andrew Siegel for helpful comments and suggestions.†Boston University. Email: [email protected]‡University of Washington. Email: [email protected]§London Business School. Email: [email protected]
1 Introduction
As part of efforts at improving government transparency and accountability, many countries
require that politicians provide financial asset disclosures on taking office.1 In some cases,
including that of India which is our focus here, public asset disclosures are required even to
stand for office.
As Djankov et al. (2010) document, compliance with asset disclosure laws is mixed. But
even if politicians’ assets are made public, there is no evidence to date on whether they have
any effect on elections or the functioning of government—the politicians that choose to stand
for office; how they behave once elected; whether disclosures affect the selection of politicians
by voters.
This paper provides, to our knowledge, the first empirical analysis of the effects of asset
disclosure laws, by examining a change in the financial disclosure requirements of Indian state-
level Members of the Legislative Assembly (MLAs). Specifically, we study the selection of
politicians around a Supreme Court ruling that has required, since November 2003, the that
all candidates standing for public office at all levels to disclose the value and composition of
their assets. Disclosure is mandatory, with punitive consequences for misreporting, and asset
disclosures are publicized via civil society organizations such as the Association for Democratic
Reforms (ADR).
For studying the impact of disclosure, the timing of the ruling was fortuitous. Its imple-
mentation date was November 2003, in the midst of the 2002-2004 wave of state elections.
As a result, 10 states held elections in the 18 months prior to the change, and 10 states held
elections in the 18 months following. We argue that a comparison of changes in pre- and post-
disclosure states allows us to credibly distinguish the effects of disclosure rules from general
time trends.
We focus on three main outcomes: politicians’ decisions to stand for public office; incum-
1See Djankov et al., 2010, for a detailed list.
2
bents’ electoral success conditional on standing for office; and asset returns of politicians that
hold office.
If effective, there are several possible ways that disclosure laws might affect politicians’
willingness to run for office. First, there might be an immediate impact, as potential can-
didates avoid disclosing their wealth, possibly because of ill-gotten rents while in office or
due to privacy concerns. But a snapshot of assets fails to provide any indication of wealth
accumulation in office, which can only be calculated from wealth disclosures across elections.
This suggests a possible lag in the effect of disclosures on politicians’ decisions to stand for
election. Furthermore, asset disclosures may change the composition of politicians that hold
office, leading to a selection of those less motivated by rent-seeking. This would predict a one-
time effect of disclosures on the decision to stand for election, and a shift in the composition
of political candidates.
Our empirical results are as follows: we find no effect on the fraction of MLAs standing for
reelection (referred to henceforth as the “re-run rate”) in the first post-RTI period – disclosures
that revealed the level, but not the growth, in assets. In the second period, however, we find
a large (13 percentage point) decline in the re-run rate. In the third post-RTI election, re-
run rates return to their pre-disclosure levels. (We emphasize that our empirical analysis
exploits differences in the timing of (pre-determined) elections across states, to compare the
trajectories of states that held elections just before the passage of the RTI Act versus those
with elections just after.)
It is difficult to distinguish a reversion to previous re-run rates from a permanent drop in
re-run rates, given the steady increase in re-run rates that preceded the RTI Act. A number
of further results lead us to favor an interpretation of the data whereby the RTI Act leads to
a one-time shift in the composition of candidates, rather than a shift toward officials holding
office for only a single term.
This interpretation is based largely on two pieces of additional analysis: First, we examine
the electoral success of MLAs that choose to stand for reelection. As noted by Linden (2004)
3
among others, standing Indian politicians have traditionally faced an incumbency disadvan-
tage (as opposed to, say, the United States, where there is a well-documented incumbency
advantage). We verify that this pattern holds within our pre-RTI sample. However, we find
that the incumbent disadvantage disappears in the second election that follows the RTI Act’s
passage, i.e., in the same election that re-run rates decline sharply. We take this as evidence
that politicians that might have been selected out by voters are self-selecting out of holding
office (for example, because ‘good’ and ‘bad’ types were pooled in the pre-disclosure era, but
could be distinguished from one another in the post-disclosure period).
Second, we find a decline in the asset growth premium of senior politicians among those
chosen by voters in this second election following the passage of the RTI Act. Specifically,
while incumbents and ministers enjoyed relatively high rates of wealth accumulation in the
early years of asset disclosure, this asset growth premium disappears in the second post-RTI
election, coincident with the decline in re-run rates.
We additionally find that, for the set of politicians who were already incumbents when
asset disclosure laws were passed (and who remain in office for the first two post-disclosure
elections), their rates of wealth accumulation decline across their two terms in office declines.
This suggests that the RTI Act had an effect on rent-seeking in office (or at least an effect on
reported wealth accumulation).
Overall, our results fit best with financial disclosures leading to a one-time shift in the
composition of political aspirants toward more electable candidates who are less likely (and
less able) to exploit their offices for private gain.
Our work contributes most directly to research on the effects of increased transparency and
accountability on the quality of government. A number of studies have used asset disclosure
data to study politicians’ wealth accumulation.2. We are, to our knowledge, the first to
look at the impact of disclosure laws on the selection and behavior of politicians, rather than
exploiting the data generated by these laws to study politicians’ wealth. A number of scholars
2See, for example, Fisman, Schulz, and Vig (2014) for an analysis of wealth accumulation of Indian MLAs;
TKSweden, TKUS; and TKUK
4
have examined how greater transparency and information disclosure affects the functioning of
government transfer programs. For example, Banerjee et al (20TK), for example, look at the
effects of providing Indonesian villagers with more information on a subsidized rice program,
while Reinikka and Svensson (20TK) examine the impact of publicizing leakage of school fund
transfers in Uganda.
We also contribute to the discussion on the determinants of politician selection and per-
formance. Ferraz and Finan (TK), TKItaly, and Fisman et al (TK), for example, examine on
the effect of bureaucratic pay on the quality of candidates, as well as their performance once
in office. Besley et al. (2013) and Banerjee and Pande (2007) consider the role of competition,
both within and across parties, while Beath et al. (TK) consider the role of electoral rules,
exploiting a field experiment in Afghanistan. We share with many of these papers an emphasis
on microeconomic identification, taking advantage of the timing of the RTI Act’s passage to
credibly identify the effects of disclosure on political selection.
2 Background and Data
2.1 Background on asset disclosure laws
Prompted by a general desire to increase transparency in the public sector, a movement for
freedom of information began during the 1990s in India. These efforts eventually resulted in
the enactment of the Right to Information Act (2005), which allows any citizen to request
information from a “public authority,” among other types of organizations. During this period,
the Association for Democratic Reforms (ADR) successfully filed public interest litigation
with the Delhi High Court requesting disclosure of the criminal, financial, and educational
backgrounds of candidates contesting state elections.3 Disclosure requirements of politicians’
wealth, education and criminal records were de facto introduced across all states beginning
with the November 2003 assembly elections in the states of Chhattisgarh, Delhi, Madhya
Pradesh, Mizoram, and Rajasthan.
3http://adrindia.org/about-adr/
5
Candidate affidavits provide a snapshot of the market value of a contestant’s assets and
liabilities at a point in time, just prior to the election when candidacy is filed. In addition to
reporting her own assets and liabilities, a candidate must disclose the wealth and liabilities
of her spouse and dependent family members. This requirement prevents simple concealment
of assets by putting them under the names of immediate family members. Further, criminal
records (past and pending cases) and education must be disclosed.
Punishment for inaccurate disclosures may include financial penalties, imprisonment for
up to six months, and disqualification from political office. While there have been a hand-
ful of revelations of politicians’ asset misstatements4 and at least one prosecution, against
Jharkhand minister Harinarayan Rai, for failing to disclose assets, for the most part pop-
ular accounts focus instead on the very high level of asset accumulation implied by these
disclosures.5
2.2 Data
2.2.1 State Assembly Election Data
The principal data on elections are collected from Statistical Reports of Assembly Elections
provided by the Election Commission of India (ECI).6 Legislative Assembly elections are held
regularly in all of India’s 28 states as well as in two Union Territories (Delhi and Puducherry)
and Members of the State Legislative Assembly (MLAs) are elected from each of the state’s
assembly constituencies (ACs) in first-past-the-post voting.
The average electoral cycle is five years. Critical to our identification strategy, elections
are staggered across states with at least some elections being held in almost every year.
Indeed, it is rare for elections to diverge from a cycle of five years. For example, all of the
states that held elections in November 2003 (Chhattisgarh/Madhya Pradesh, NCT of Delhi,
4For example, Firstpost India reported that Himachal Pradesh MLA Anil Kumar failed to declare ownership
of a pair of properties in his 2007 disclosure.5See, for example, “How the political class has looted India,” The Hindu, July 30, 2012,
Rajasthan, Mizoram) also held elections in the same month 5 years earlier (November 1998).
The same holds for the states with elections February 2003 (Himachal Pradesh, Meghalaya,
Nagaland, Tripura) which all had previous elections in February 1998. This indicates that the
sorting of elections around the passage of disclosure requirements is unlikely to be an issue.
(Additionally, election dates are set well in advance, making it all the more likely that sorting
would be a concern.)
For each state and union territory, we collect data from all available reports beginning
up to five elections prior to the first election with mandated disclosure of candidate affidavits
(henceforth referred to as election e(1)). Table 1 provides an overview of the state assembly
elections in our sample along with some general descriptive statistics. For 13 of the 30 states,
we observe three elections post implementation of disclosure requirements.
Overall, the data consist of 30,398 assembly constituency elections, comprising a total of
299,967 candidate observations with information on candidate name, gender, party, and vote
outcome, as well as information on constituency-level reservation status (Scheduled Caste
(SC), Scheduled Tribe (ST), or “general”)7, voter turnout, and electorate (for elections post
2003, reports also include candidate age and caste category (i.e., scheduled caste, scheduled
tribe, or general). On average, each constituency covers an electorate of about 145,000. Voter
turnout in ACs averages about 65.26 percent (standard deviation of 13.61 percent) and just
less than five percent of candidates are women.
Matching Candidates: For each assembly constituency election, we match winners and
runners-up with candidates that contest in the subsequent election of that constituency. In a
first step, we employ a fuzzy matching algorithm that, among other things, accounts for dif-
ferential spelling of names across elections. Due to the many commonalities across names, in
the second step we manually check the set of all suggested matches, discarding those matches
that prove unlikely to be the same candidate. For example, “A.R.KRISHNAMURTHY”
7SC and ST constituencies are reserved for candidates classified as SC or ST, in accordance with a policy
introduced to promote the representation of historically under-represented groups; general candidates cannot
compete in these constituencies.
7
(Santhemarahalli AC in Karnataka election 1999) is unlikely to be the same candidate as
“KRISHNA MURTHY MS” in the subsequent election (and in fact, verifiably distinct). On
the other hand, “RATHOD ANIL (BHAIYYA) RAMKISAN” and “ANILBHAIYYA RAMK-
ISAN RATHOD (B.COM)” (Ahmednagar South constituency in Maharashtra elections 1999
and 2004) are deemed a likely match even though the names in the ECI reports are different.
. Five smaller states experienced reorganizations after elections during the 1980s, which
resulted in changes in the number and naming of constituencies (for example, Arunachal
Pradesh had 30 ACs in the 1984 election and 60 ACs in the 1990 election). We do not
attempt to match candidates in those years.
Asset disclosure requirements commenced with the November 2003 state elections and
all assembly elections had mandatory disclosure of candidate affidavits by 2008. While we
match candidates within constituencies prior to disclosure, post disclosure matching is done
within state, that is, we account for politicians that choose to rerun but switch constituencies
within a state across elections. This is largely necessitated by renumbering and renaming of
constituencies between elections post 2003. Note that this approach allows for a consistent
comparison of re-run probabilities of candidates at e(0) – the last election prior to disclosure
– with re-run probabilities of contestants at e(-1) and earlier.
This approach may cause an upward bias when comparing re-run probabilities of candi-
dates at e(1) – the first election with disclosure – with re-run probabilities of contestants at
e(0), since within-state matching is more likely to generate a candidate match than within-
constituency matching. Given this upward bias, we argue that our estimates of asset disclosure
on re-run propensity (which is negative) are plausibly biased toward zero. (This approach
also alleviates possible concerns of increased labor mobility over time and within state that
would otherwise not be accounted for in within-constituency matching.8)
For candidate i in state s that stood for election at time t, we define the indicator variable
Rerunist to denote whether i was also a candidate in the next election. We define the state-
8We further verify that across-state mobility is virtually non-existent, that is, politicians are state-bound.
8
election level variable Disclosurest to denote whether asset disclosures are required at time
t. Our main interest will be in studying the Rerun probabilities of MLAs. In a set of placebo
regressions, we will examine the decisions of politicians that stood for office at t but came in
second (the “runners-up” sample) choose to run in the next election.
Over the entire sample of assembly constituency elections, winners on average re-run 72.22
percent of the time while runners-up re-run 42.11 percent of the time.
2.3 Candidate Affidavits
National Election Watch, in collaboration with ADR, provides digitized records based on
affidavits from candidates in state assembly elections. These records serve as the basis for
most of our data on asset disclosure. For nine elections in our sample prior to October 2004,
however, digitized records were unavailable, so we collected data directly from scanned pictures
or PDFs of candidate affidavits. The affidavits were gathered from either the GENESYS
Archives of the Election Commission of India (ECI)9 or the various websites of the Office of
the Chief Electoral Officer in each state. Overall, there is affidavit information for 73 elections,
and 13 of the 30 states in our sample have three elections with asset disclosure. A sample
affidavit is shown in the Appendix.
Since reporting requirements are limited to those standing for election, asset growth can
only be measured for re-contesting candidates, i.e., those that contest and hence file affidavits,
in two elections. For the sample of matched candidates, we filter out affidavits that are poorly
scanned, have missing pages, or handwriting that is too unclear or ambiguous to get a clear
picture of a candidate’s reported financial situation. Next, we verify suspicious values and,
since a main focus is on growth in wealth, remove candidates that list significant assets without
corresponding market value information, leaving a sample of 5,352 matched candidates (3,403
winners and 1,949 runners-up). From the affidavits, we compute each candidate’s Net Wealth
at the time of filing, just prior to each election. In each case, we define net wealth as the
sum of movable assets (such as cash, deposits in bank accounts, and bonds or shares in
9http://eci.nic.in/archive/
9
companies) and immovable assets (such as land and buildings) less liabilities (such as loans
from banks), aggregated over all dependent family members listed on the affidavit. For the
5,352 candidates, we have 1,214 constituency elections in which both the winner and the
runner-up re-contest in the following election. We define Final Net Wealth as net wealth at
the end of the electoral cycle under consideration, and Initial Net Wealth as net wealth at
the beginning of the cycle.10
We also generate a number of control variables from the candidate affidavits. We define a
Criminal Record dummy as equal to one if the candidate has pending or past criminal cases
at the time of the first election, and measure education based on years of schooling (Years of
Education). We also distinguish among winning candidates based on whether they went on
to hold significant positions in the state government, using an indicator variable Minister to
denote membership in the Council of Ministers, the state legislature’s cabinet.11 Table 2 lists
definitions of the main variables used in the analysis.
3 Empirical Strategy
In this section, we lay out our basic estimating equation, and discuss our approach to identify-
ing the effect of disclosure, as distinct from time trends. As noted in the preceding section, the
precise timing of elections is critical to making this distinction. If all state assembly elections
occurred concurrently, it would be impossible to distinguish disclosure’s effects from the time
trends that are evident in the data.
As noted above, five states held elections concurrent with the advent of asset disclosure
requirements (Chhattisgarh, Delhi, Madhya Pradesh, Mizoram, and Rajasthan). In just the
eight months preceding November 2003, four other states held elections. In all of these cases,
the election schedule was set well before the timing of disclosure requirements – which were
created as a result of a court ruling – was apparent. More broadly, in the 36 months around the
10For further details, see Fisman, Schulz, and Vig (2014).11Most information was sourced from archives of state government websites as well as an extensive review
of newspaper articles.
10
November 2003 implementation of disclosure requirements, a total of 20 states held elections,
10 in the 18 months preceding this date (“just before states”) while 10 states held elections
in the 18 months that followed (“just after states”). See Table 1 for details on timing.
Table 3 compares the basic attributes of just before versus just after states – we observe
no significant differences between the two in terms of size, literacy, income, corruption (as
measured by Transparency International’s state-level ranking), population, scheduled caste
concentration, or voter turnout. This lack of unobserved differences between the two groups
lends some credibility to the claim that sorting around the implementation of asset disclosure
requirements is essentially random.
Our basic estimating equation exploits this difference in timing of elections pre- and post-
disclosure to separate disclosure effects from time trends. The basic intuition of our analysis is
captured in Table 4, where we show the Rerun probability of MLAs as a function of elections
relative to the advent of asset disclosures. Focusing on elections immediately around the
advent of disclosure requirements, we observe an increase in rerun probability between e(-2)
and e(-1) for the just before subsample, which span the years 1993 and 1997. For the just
after subsample over approximately the same time period (1993 - 1998), we similarly observe
a small increase in the rerun rate. This suggests some common time trend between the two
groups. However, in 2003 the two sets of states diverge – for politicians in the just before
subsample elected in 2002-2003 at e(0), the probability of standing for reelection continues to
increase. However, for those elected in just after states in 2003-2004 at e(1), there is a steep
drop in rerun probability. Interestingly, one election cycle later, MLAs in just after states
experience a drop in rerun probability. The fact that the drop in rerun probability appears to
be timed to election cycles relative to disclosure requirements, rather than timed to election
date, is the basis of our claim of a causal effect of disclosure.
In Appendix Table A-1 we show a comparable table to examine the rerun probabilities
for runner-up candidates. Interestingly, we observe no drop in their odds of recontesting,
indicating that the drop in rerun rates for MLAs associated with disclosure does not reflect a
11
general decline in interest in running for office.
Before proceeding to our main specification, it is worth noting that Table 4 also shows
some divergences in other election cycles between the two subsamples. This adds noise to the
identification of a post-disclosure drop in Rerun.
The direct effect of Winner captures the incumbent (dis)advantage in an election where
disclosure will not be required in the next election. The interaction term Winner∗Disclosure
captures the change in incumbency advantage that comes with disclosure.
We present the results in Table 6. Column (1) indicates a pre-disclosure incumbent dis-
advantage of 5.7 percent (significant at the 10 percent level), comparable to estimates from
Linden (2004). Then interaction term, Winner ∗ Disclosure, has a coefficient of 0.101, in-
dicating that incumbency switches (weakly) to being an electoral advantage after the advent
of disclosure requirements. The addition of party (column (2)) and party*state fixed effects
(column (3)) has little effect on these point estimates; similarly, the inclusion of a range of
controls (column (4)) has very little effect on the estimated incumbency effect, and how it is
affected by disclosure. In columns (5) and (6) we limit the sample to close elections, those
won by 10 and 5 percent respectively. Not surprisingly, the pre-disclosure incumbency disad-
vantage is far stronger in relatively close elections, but in column (5) the coefficient on the
interaction term Winner ∗ Disclosure is largely unchanged. In column (6), the interaction
term loses its significance (p-value of 0.11), and shrinks in magnitude by about a third.
4.3 Behavior in office
We finally consider the behavior of politicians who do choose to continue running for office,
and are elected by voters to do so. We focus on wealth accumulation as the most directly
impacted measure of MLA “performance.” Specifically, we examine whether, holding initial
wealth constant, election winners’ assets grow more rapidly than the corresponding sample of
14
election runners-up.
In earlier work, we showed that, among a sample of winners and runners-up in state
assembly elections who choose both to recontest in the next election (essential to calculating
asset growth), wealth accumulation is higher for winners. This “winner’s premium” is greater
for higher-level politicians, most notably among ministers and reelected incumbents. In these
analyses, we could examine only a single period in office (2004 - 2009). Using more recent
data, we can compare asset growth for the first and second post-disclosure periods.
A pair of major caveats are in order in interpreting this section’s results. First, whereas
we could exploit small differences in the timing of elections around the implementation of
disclosure requirements to distinguish disclosure effects from time effects, this is not possible
in comparing asset growth in the two post-disclosure periods. Since we require 10 years of
post-disclosure data to study two periods of asset growth, our estimate asset growth during
the second post-disclosure period comes from only those states that held elections shortly
after disclosure implementation. Second, we are measuring the asset returns of politicians
based on their public disclosures. We might observe a change over time because of a shift in
rent seeking, or simply because politicians become more adept at hiding their assets.
In Table 7, columns (1) and (2), we compare the “winner’s premium” for the period e(1) -
e(2) versus the period e(2) - e(3) respectively. The coefficient on Winner is virtually identical
across the two periods. (While column (1) includes all states, column (2) only includes those
which have held at least three post-disclosure elections. However, we observe near-identical
patterns in this Table and the one that follows if we limit the e(1) - e(2) sample to those states
that held at least three post-disclosure elections.) Columns (3) and (4) show similar patterns
in a pooled regression across both periods - that is, on average the stated asset returns of
politicians is unchanged by disclosure.
However, this similarity in averages masks considerable heterogeneity in the asset growth of
different types of politicians. In particular, in earlier work we show that the winner’s premium
in asset growth increases as MLAs progress through the state-level political hierarchy: the
15
winner’s premium is higher for incumbents than freshmen, and yet higher for ministers.
In Table 8 we observe that this pattern changes in the second period after disclosure
requirements were put in place. Comparing columns (1) and (2), we find that the higher
winner’s premium associated with incumbency drops to zero in the second post-disclosure
period (column (2)). At the same time, the winner’s premium for non-incumbents increases
from near zero to about 17 percent. We observe a similar pattern for ministers in columns
(4) and (5), though the difference between the minister’s premium across the two periods is
not significant at conventional levels (see column (6)).
Finally, in Table 9 we look at the returns of candidates who held office during both periods
e(1) - e(2) and e(2) - e(3), as compared to runner-up wealth accumulation. By definition, these
are all MLAs who won at e(2) as incumbents so that, based on previous estimates of asset
growth, we might expect higher asset growth in the later period. However, we find instead a
decrease in the winner’s premium for this consistent set of 331 MLAs who held office across
the two election cycles that followed the implementation of disclosure rules.
DISCUSSION TO BE ADDED
5 Conclusion
16
References
Cameron, A. Colin, Jonah B. Gelbach and Douglas L. Miller, 2008, Bootstrap-Based Improvements
for Inference with Clustered Errors, The Review of Economics and Statistics 90(3), 414-427.
Djankov, Simeon, Rafael La Porta, Florencio Lopez-de-Silanes and Andrei Shleifer, 2010, Disclo-
sure by Politicians, American Economic Journal: Applied Economics 2(2), 179-209.
Fisman, Raymond, Florian Schulz and Vikrant Vig, 2014, The Private Returns to Public Office,
Journal of Political Economy 122(4): 806-862.
Linden, Leigh, 2004, Are Incumbents Always Advantaged? The Preference for Non-Incumbents in
India, Working Paper.
17
Table 1: Overview of State Assembly Elections
Notes: This Table provides an overview of the state assembly elections in our sample along with some general descriptive statistics (this data corresponds
to the gray-shaded election years). Election e(1) indicates the first election post the information disclosure reform of 2003. Panel (A) lists states that had
elections immediately following the disclosure reform (2003/04) and Panel (B) lists the subset of states that had elections just prior to the reform (2002/03).
Panel (C) list the remaining states in our sample. *Carved out of Madhya Pradesh; **carved out of Uttar Pradesh; ***carved out of Bihar (all in 2000).
Source: Statistical Reports on General Elections, Election Commission of India, New Delhi (various years).
Notes: This Table investigates the effect of the disclosure reform on subsequent electoral success of re-contesting candidates. This sample consists of
paired constituency winners and runners-up of the “just post” and “just prior” states shown in Table 1. Disclosure is an indicator that is defined as 1 if
recontesting will require the disclosure of subsequent affidavits (which allows to measure wealth accumulation over the election cycle). All specifications
control for general time trends. Bootstrapped standard errors clustered at the state-level are given in parentheses. Coefficients with ***, **, and * are
statistically significant at the 1%, 5%, and 10% levels, respectively.