AN EXAMINATION OF STOCK OWNERSHIP AND WEALTH ACCUMULATION
BY U.S. CONGRESS MEMBERS
By
Corrine Carr

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Business Administration—Finance—Doctor of Philosophy
2017

ABSTRACT
AN EXAMINATION OF STOCK OWNERSHIP AND WEALTH ACCUMULATION
BY U.S. CONGRESS MEMBERS
By
Corrine Carr
Given the profound impact actions of the federal government have on financial markets,
U.S. congressmen should be capable of enhancing their wealth through investing on their
private government information. However, existing research on U.S. congressmen’s common
stock trading has yielded conflicting results on whether they trade on their private information.
In essay 1 of this dissertation, I reexamine whether members of Congress traded on private
information from 2004-2012, and conclude that in the aggregate, they do not. However, House
members perform extraordinarily well, earning abnormal returns of 8.2% for a 12-month
holding period during their final term in office. This is likely due to incentives provided by the
House Committee on Ethics and disinterest in reelection. Essay 2 builds on the first essay by
focusing on returns to congressmen’s total wealth, which captures their gains in the opaque
private and real estate markets where they are more likely to act corruptly, in addition to their
returns in public markets. I find that members on average earn abnormal returns to their
wealth of 6.39% per year. Using the models suggested by Treynor and Mazuy (1966) and
Henriksson and Merton (1981), I find that the abnormal returns to wealth are due to selectivity
rather than market timing. Overall, the dissertation suggests that public financial disclosure is
effective and reconfirms that incentives are important drivers of behavior. It also supports the
public perception that congressmen unethically trade on private government information.

ACKNOWLEDGEMENTS

I am highly indebted and enormously grateful to Dr. Andrei Simonov. My main regret over the
past few years is that I didn’t start working with him earlier in my academic career. His expert
guidance, understanding and support helped me through many ups and downs. I am thankful
for the interest he has shown in my research and the advice regarding the topic of my research.
Thank you for chairing my committee.
I am grateful to Dr. Charles Hadlock for his encouragement, motivation, academic and practical
advice. His excitement for academia is inspiring. I truly enjoyed my time in your class. Thank
you for the knowledge you passed on.
I would like to express my appreciation to Dr. Geoff Booth and Dr. Zoran Ivkovic for serving on
my committee. Their thoughtful questions and comments were valued greatly.
I would also like to thank Dr. Miriam Schwartz-Ziv. It is because of her that I stumbled upon my
dissertation topic.
Finally, thanks to my family, friends, and fellow graduate students for all of the encouragement
and support during the program.

iii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................................................................v
CHAPTER 1: Introduction ................................................................................................................ 1
CHAPTER 2: Stock Ownership by U.S. Congress Members ............................................................. 4
2.1 Introduction .................................................................................................................. 4
2.2 Congressional Rules of Conduct.................................................................................... 7
2.2.1 House of Representatives ................................................................................... 8
2.2.2 The Senate .......................................................................................................... 9
2.3 Theoretical Framework ............................................................................................... 10
2.4 Data Description.......................................................................................................... 12
2.5 The Performance of Congressmen’s Trades ................................................................ 14
2.6 Congressmen Net Worth............................................................................................. 17
2.7 The Performance of Stocks Traded by Congressmen Close to Retirement................. 20
2.8 Alternative Explanations for Abnormal Returns ......................................................... 24
2.8.1 Net Worth ......................................................................................................... 24
2.8.2 Seniority............................................................................................................ 28
2.9 Conclusion ................................................................................................................... 31
CHAPTER 3: Wealth Accumulation by U.S. Congress Members ................................................... 33
3.1 Introduction ................................................................................................................ 33
3.2 Data Description.......................................................................................................... 38
3.3 Methodology ............................................................................................................... 41
3.4 Empirical Results ......................................................................................................... 45
3.5 Powerful Congressmen ............................................................................................... 51
3.6 Conclusion ................................................................................................................... 53
BIBLIOGRAPHY .............................................................................................................................. 56

iv

LIST OF TABLES

Table 2.1: Summary of Annual Transactions ................................................................................ 13
Table 2.2: Performance of Stock Traded by Members of Congress ............................................. 15
Table 2.3: House and Senate Net Worth (Top 15 annual % Increase) ......................................... 19
Table 2.4: Raw Returns and DGTW adjusted returns 1-8 years before retirement ..................... 22
Table 2.5: Summary of Congressmen Net Worth ........................................................................ 25
Table 2.6: Raw and DGTW Adjusted Returns for Net Worth Quintiles ....................................... 26
Table 2.7: DGTW Adjusted Returns for Various Seniorities......................................................... 29
Table 3.1: Summary of Net Worth and Assets .............................................................................. 40
Table 3.2: Fama-French Three Factor Model................................................................................ 46
Table 3.3: Treynor-Mazuy and Henriksson-Merton Models ........................................................ 47
Table 3.4: HCAPM Model .............................................................................................................. 49
Table 3.5: Returns to Wealth for Powerful Members of Congress .............................................. 52

v

CHAPTER 1: Introduction
In recent decades large corporations have become increasingly involved in politics.
Today, many corporations devote enormous resources to lobbying and making donations to
politicians. This increase in corporations participating in government affairs and the
development of relationships between corporate executives and politicians has generated
concerns about whether corporations are acting in the best interests of their shareholders,
whether politicians are acting in the best interests of their constituents, and the ramifications
this could have on the U.S. economy. These concerns have resulted in an abundance of
research and news reports addressing these questions, numerous shareholder proposals
requesting that corporations disclose their political contributions, and the 2012 STOCK Act
which prohibits congressmen from using non-public information gained through their official
duties for personal profit.
The research that has blossomed from these worries over the relationships between
corporations and politicians suggests there are reasons for concern. That is, these studies
support the hypothesis that corporations benefit from these relationships, making a positive
return on their investments in politics, and suggest this may be a negative for the economy.
Most recently, James Bessen (2016) examines whether rising corporate profits are the result of
profitable, innovative investments or political rent seeking and finds that, since 2000, the rise in
profits has been largely due to political activity and regulation. These results raise concerns
about the effects political rent seeking could have on the economy. The link between
regulation and profits is concentrated in only a few industries, including pharmaceuticals/
chemicals, petroleum refining, transportation equipment/defense, utilities, and
1

communications. Such a concentration of political influence could impact policy for the entire
economy by skewing it towards these few industries. Additional concern is caused by the
enormous amount of financing currently going towards influencing politicians, and the growth
in the strength of the relationship between the political and corporate worlds. Since 2000,
political campaign spending by firm Political Action Committees (PACs) has increased by thirty
times and the Regdata index, which quantifies federal regulation, has increased by
approximately 50% for public firms.
Changes in where many corporations are directing investments and the potential
impacts on the economy warrant investigations into whether politicians also benefit from these
relationships. If politicians and corporations mutually benefit from their relationships with each
other, then politicians may be less willing to regulate the flow of money from corporations to
Washington, if this is indeed needed. Existing research suggests congressmen benefit from
these relationships through increased contributions to their campaigns. There have been
mixed results on whether U.S. congressmen make gains in other ways.
Currently, to discourage congressmen from profiting on private government information
they are required to file annual reports disclosing detailed information on their financial
holdings and transactions in income-producing property, according to the “Ethics in
Government Act of 1978”. How effective are these public disclosures in discouraging unethical
behavior? In this dissertation, I address this question. In chapter 2, I examine whether
congressmen earn positive abnormal returns on their stock transactions. If they do, then they
likely trade on private information. In chapter 3, I examine whether congressmen earn
abnormal returns to their total wealth. While public disclosure may deter congressmen from
2

trading on their private information in public equity markets, it may not deter them from using
their legislative power and private information to make gains in more private markets such as
real estate and private business.

3

CHAPTER 2: Stock Ownership by U.S. Congress Members
2.1 Introduction
In recent years, members of Congress have received extensive criticism for investing on
private government information. There is no shortage of anecdotal evidence that congressmen
may be able to take advantage of their positions in government to make a nice living outside of
their standard government salary. The speculation ranges from Senator John Kerry’s well timed
trades in pharmaceutical companies around the development and passing of the Affordable
Care Act to Representative Nancy Pelosi’s impressive track record for IPO participation. The
most prominent criticism has come from Schweizer (2011), who provides a plethora of
anecdotes of suspiciously well timed stock trades, which resulted in enormous profits, and
Ziabrowski et al. (2004, 2011) who find that the average Senator beats the market by 12% per
year, while House members beat the market by 6% per year. To put these returns in
perspective, the average corporate insider beats the market by about 7% a year (Jeng, Metrick,
and Zeckhauser 2003). In a recent paper, Eggers and Hainmueller (2013), question the results
found by Ziabrowski et al., which show that congressmen earn large abnormal returns, by
finding that congressional portfolios, rather than their transactions, show no evidence of
superior stock picking ability or positive abnormal returns for the years 2004 to 2008.
The findings of Eggers and Hainmueller (2013) are surprising given that much research in
political economy suggests that firms and politicians benefit mutually from their relationships
with each other. There are many examples of the financial benefits of political connections to
firms. It has been found that political connections help firms gain access to bank loans (Khwaja
and Mian, 2007; Faccio, 2006), receive more government contracts (Tahoun, 2014), increase
4

firm value (Roberts, 1990; Fisman, 2001) or improve firm performance (Johnson and Mitton,
2003). In addition to these benefits of politician-firm connections, Yu and Yu (2006) find firms
that lobby can evade fraud detection on average 117 days, and are 38% less likely to be
detected by regulators than firms that do not lobby. There is also evidence of the financial
benefits to politicians through their connections to firms. Tahoun (2014) finds that U.S.
congressmen tend to invest in firms that make large contributions to their campaigns, thus
providing an incentive for them to act in these firms’ best interests.
The purpose of this chapter is to reexamine whether congressmen’s stock investments
tend to outperform the overall market and to examine the effects of self-imposed term limits
through retirement on congressmen’s behavior. Finding that they outperform the market
would support the hypothesis, and popular perception, that congressmen use private
government information for personal gain. Eggers and Hainmueller (2014) find that U.S.
congressmen earned mediocre returns from 2004 to 2008 because of their “non-connected”
investments, but their investments in local firms and campaign contributors beat the market.
The mediocre returns during this time period could also be a result of the recent increased
attention on congressmen taking advantage of their positions in government for personal gain.
Concerns about this possibility led to the proposal of the STOCK Act in 2006 and its passage into
law in 2012. The STOCK Act made it illegal for members and employees of Congress to profit
from nonpublic information they obtain from their government positions, although they may
trade on information they gain through their legislative roles as long it doesn’t come from

5

private meetings and involve confidential information.1 There has also been increased
attention on firm-politician connections, resulting in an increasing number of shareholder
proposals suggesting that firms disclose details of their political activities.
The recent attention on firm-politician relationships may deter congressmen from
trading on nonpublic government information in the aggregate. However, situations may arise
in which the gains of trading on private information outweigh the risks. Specifically, the
benefits may outweigh the costs when a House member is already planning to leave office,
effectively self-imposing a term limit. Since the House Committee on Ethics doesn’t have
jurisdiction over former members, departure from the House ends an ethics investigation. This
is best explained by a New York Times article on the departure of Michele Bachman, Tim
Bishop, Tom Petri, Paul Broun, and Steve Stockman from the House in 2014:
As the lame duck Congress plays out its final days, five departing House
members are breathing sighs of relief. They have been the focus of longrunning ethics investigations, but their cases will depart with them from the
public record, closed without any official findings. The slates will be instantly
cleaned under a provision of the often arcane House ethics rules, which critics
complain encourages procrastination and shady behavior.
Thus, if a member of Congress is planning to leave office, she may take advantage of
nonpublic information to earn abnormal returns in the stock market without
repercussion.
1

Congress attempted to block the Securities and Exchange Commission’s first major investigation of political
insider trading. The House refused to hand over evidence and claimed Brian Sutter, the staff member under
investigation, had legal immunity. This investigation has not resulted in formal allegations of insider trading.
(Mullins, B. and Ackerman, A. (2015, November 17). Court says Congress must comply with Federal insider-trading
investigation. The Wall Street Journal. Retrieved from http://www.wsj.com)

6

It has been well documented in the political agency literature that lame duck politicians,
those who face term limits and are currently in their final term, tend to exert less effort (Besley
and Case, 1995; Alt, de Mesquita, and Rose, 2009). Elections are an important instrument
through which voters can hold politicians accountable and thus mitigate moral hazard.
Politicians will act in the best interests of the voters if they know they will only be reelected if
they perform better than some standard set by the voters (Barro 1973; Ferejohn 1986). Most
of the political agency literature is focused on whether retiring members exert less effort.
However, some research has focused on whether politicians act corruptly during their final
term when they are term limited. Ferraz and Finan (2011) investigate the level of corruption
during Brazilian mayors’ final term and find that those who are eligible for reelection engage in
less corruption, on average, than term-limited mayors. They also find that voters tend to
punish corrupt politicians when their corrupt practices are publicized (Ferraz and Finan, 2008).
This chapter contributes to the literature by focusing on whether retiring members act in their
own best interests through corruption, and by showing that moral hazard problems may exist
with or without term limits.
2.2 Congressional Rules of Conduct
The Constitution grants Congress the authority to discipline its members. This authority
is found in Article I, Section 5, clause 2 of the Constitution, which explains “each House may
determine the rules of its proceedings, punish its members for disorderly behavior, and, with
the concurrence of two thirds, expel a member.”2 Since 1967, both the House and Senate have

2

U.S Congress, House, “Article I, Section 5, clause 2,” The Constitution of the United States, 108th Cong., 1st sess.,
H.Doc. 108-96 (Washington: GPO,2003), p.4.

7

had permanent ethics committees, which were eventually called the House Committee on
Ethics and the Senate Committee on Ethics. These committees were established to create
formal rules of conduct and disciplinary procedures to investigate and punish illegal or
unethical acts by their members.3 While the Senate and House Committees on Ethics have
similar goals, there are significant differences in their ability to enforce ethical behavior.
2.2.1 House of Representatives
Currently, the House Committee on Ethics is comprised of ten members, five from each
party. Accusations of misconduct or violations of House rules by House Members or staff can
only be filed with the Committee on Ethics by a Member of the House. Accusations may be
filed by nonmembers only if it is accompanied by written certification by a Member that the
accusations warrant review by the committee. In 2008, the Office of Congressional Ethics (OCE)
was established to collect information from non-Members, thus creating an external body to
review complaints against House Members. If the Committee on Ethics decides that an
investigation is necessary, an investigative subcommittee is formed to judge the evidence of
ethics violations. Following a subcommittee investigation, the Committee on Ethics has
historically recommended punishments including expulsion, censure, reprimand, and “letters of
Reproval” and “Letters of Admonition”. About 25 House Members have left the House (by
resigning, not running for reelection, or defeat) after court convictions, after inquiries were
initiated, or after charges were brought by the committee.4 This departure from the House
ends an investigation because the committee does not have jurisdiction over former Members.
3

Straus, J.R., 2011, Enforcement of congressional rules of conduct: An historical overview. In Congressional
Research Service, Washington, DC, June (Vol. 14).
4
Maskell, J., 2002, Expulsion, censure, reprimand, and fine: legislative discipline in the House of Representatives.
Congressional Research Service

8

This greatly reduces the power of the House Committee on Ethics and is one of the greatest
differences between the House Committee on Ethics and the Senate Committee on Ethics.
2.2.2 The Senate
The Senate Committee on Ethics consists of six members, three from each party. Unlike
the House committee, the Senate committee has no “statute of limitations” for investigations
of past violations. This means Senators can be charged with a crime any number of years
following their time in office. Also, there are no restrictions on who can file an allegation with
the committee. If the committee receives a complaint, it begins preliminary inquiry, and if it
finds substantial evidence of a violation, it determines the appropriate charges. Historically, as
a result of committee action, Senators have been expelled and censured for their behavior.
Many expulsion proceedings have been initiated by the Senate that did not result in expulsion.
In most of these cases the expulsion failed to obtain the necessary two-thirds vote in the
Senate. There have also been cases when the Senator facing expulsion resigned. Most
recently, in 2011 Senator John Ensign faced expulsion for improper financial conduct stemming
from his extramarital affair. He resigned before the expulsion was voted on. 5
In addition to expulsion and censure, it is possible to punish a Senator with a fine,
imprisonment, or suspension of privileges. These punishments are far less common. The latest
case was Senator Harrison A. Williams Jr. in 1981. He was convicted on nine counts of bribery
and conspiracy, fined $50,000, and sentenced to three years in prison.6

5

Lipton, E. (2011, April 21). Senator Ensign to Resign Amid Inquiry. The New York Times. Retrieved from
http://www.nytimes.com
6
U.S. Senate Historical Office, United States Senate Election, Expulsion and Censure Cases: 1793-1990 (Washington:
Government Printing Office, 1995), pp. 434-437.

9

2.3 Theoretical Framework
It is often believed that the desire to maintain a good reputation is what keeps politicians
in check. Politicians that wish to run for reelection in the future must act in the voters’
interests to win their votes. I hypothesize that an incumbent who wishes to run for reelection
will try to maintain a good reputation, and therefore will choose a low level of corruption.
However, if the incumbent decides to retire, and thus decides she does not care about
reelection, then she will choose a high level of corruption.
The objective of the following model is to show that a term limit, even a self-imposed
term limit, may have implications for a politician to choose to be corrupt. I consider a model of
political agency, since voters and politicians engage in a principal-agent relationship (Barro
1973). The interests of voters and politicians are not perfectly aligned, thus the authority given
to politicians may cause them to take actions in their own best interests rather than taking
actions that are in the best interest of the voters.
I assume the incumbent politician earns rents by acting corruptly, and that this
corruption reduces voter utility. The incumbent’s objective is to maximize the sum of current
and future rents. The preferences of the incumbent can be written as
𝑈(𝑟) = 𝑟(𝜅) + 𝑝(𝜅)𝛿𝑉
where 𝑈 is the incumbent’s expected utility. This depends positively on current rents, r, and
expected future rents, 𝑝(𝜅)𝛿𝑉, where 𝑝(𝜅) is the probability of being reelected and 𝛿𝑉 is the
politician’s discounted continuation value, which reflects her expected future utility if in office.

10

I assume the probability of being reelected depends negatively on the amount of corruption the
politician engages in. The variable 𝜅 measures corruption.
The intuition of the model is straightforward. The incumbent gains utility by acting
corruptly and the voter responds to this by conditioning her vote on the level of corruption. If
the level of corruption is too high, the voter will punish the incumbent by voting her out of
office. Thus, the incumbent faces a trade-off. A higher level of corruption reduces her
probability of reelection. As a result, to maximize her utility the politician chooses the level of
corruption that makes the voter indifferent about reelection. She will choose this level of
corruption for as long as she cares about reelection.
Now consider a term limited incumbent. That is, the incumbent is certain not to have a
next period in office. In this case the probability of reelection is zero, 𝑝 = 0, regardless of r, or
the level of corruption. The incumbent does not lose anything in terms of reelection
possibilities by acting corruptly. Thus, the incumbent’s new utility function is given by
𝑈(𝑟) = 𝑟(𝜅). This is consistent with Besley and Case’s (1995) finding that term limited
governors tend to care less about economic policy than governors who have a reputation to
uphold.
The model suggests that if the threat of being voted out of office effectively keeps
politicians in check, then there should be little or no evidence of corruption until a politician’s
final years in office. Thus, I expect that congressmen earn insignificant abnormal returns on
their investments until they decide to retire, suggesting that they do not invest on private
information until there is no longer a threat of not being reelected.
11

2.4 Data Description
My data on politicians’ equity transactions comes from annual financial disclosure forms
submitted between 2004 and 2012 by members of Congress and transcribed by the Center for
Responsive Politics. According to the “Ethics in Government Act of 1978”, each year by March
15, members of Congress are required to report their common stock purchases or sales, the
date of the transactions, and the approximate value of the transactions. This7 data has some
limitations. First, none of the financial disclosure reports (FDRs) are audited for accuracy.
Although these FDRs are not audited, congressmen may not knowingly and willfully make false
statements on their FDRs according to the “False Statements Accountability Act of 1996”. 8
Second, the care used to fill out the reports varies widely across congressmen. Some are
handwritten, some include monthly financial statements from their brokerage firms, and some
do not provide specific dates of their transaction (instead they may write that they traded a
company “4 times” or “monthly”). Third, the data does not allow for a specific measure of the
profits earned by individual congressmen. They are required to report the dollar value of
transactions within broad ranges of $1,001 to $15,000, $15,001 to $50,000, $50,001 to
$100,000, $100,001 to $250,000, $250,001 to $500,000, $500,001 to $1,000,000 and over
$1,000,000.
My dataset includes common stock transactions made by the members of Congress.
The transactions are recorded with the name of the congressman, the transaction date,

7

While members of Congress are free to invest in blind trusts, which block the owners from active involvement in
their investments, very few congressmen choose to do so.
8
Representative Charles Rangel was censured in 2010 for improperly leasing rent-controlled apartments, failing to
disclose rental income on his personal financial disclosure forms, and failing to report taxes on rental income,
among other ethics charges

12

Table 2.1: Summary of Annual Transactions
Panel A:
HOUSE
Min
25th pctile
Median
75th pctile
Max
Mean

BUYS
$ Value
$53
8,001
8,001
21,522
3,000,001
37,896

Panel B:
SENATE
Min
25th pctile
Median
75th pctile
Max
Mean

SELLS
Number
1
2
5
15
2175
43

$ Value
$498
8,001
8,001
26,375
11,700,000
59,003

BUYS
$ Value
324
8,001
8,766
32,500
1,000,001
51,062

Number
1
2
5
12
1113
26
SELLS

Number
1
3
8
37
497
35

$ Value
324
8,001
11,501
37,681
10,200,000
109,060

Number
1
2
6
24
379
26

Summary statistics are annual averages across the 2004-2012 period based on end of year financial disclosure reports for the 477
members of Congress that report common stock transactions between 2004-2012. These averages are conditional on the
congressmen trading in common stock.

whether the transaction was a buy or sell, and the approximate value of the transaction. The
common stocks were matched by hand with CUSIP numbers from the Center for Research in
Security Prices (CRSP) databases. Overall, the dataset includes 84,938 reported transactions in
a total of 2,629 companies by 477 members of congress. During 2004-2012, only 477 of the
888 congressmen who were in office invested in common stock.
Table 1 presents summary statistics describing the stock transactions of members in my
dataset. For each member who invested in common stocks from 2004-2012, I calculate the
value and number of transactions in each year and then average across years. As in Ziabrowski
et al. (2004), Ziabrowski et al. (2011), and Eggers and Hainmueller(2013), the distribution of
annual transactions is right skewed.
13

2.5 The Performance of Congressmen’s Trades
If congressmen have superior stock picking skills or trade on nonpublic information, then
stocks recently purchased should outperform their benchmarks, while stock recently sold
should not outperform their benchmarks. However, if the average congressman does not trade
on private information then I should find no relation between stock returns and congressmen’s
trades. Focusing on transactions by congressmen should provide a more powerful test of
whether they trade on private information than examining the performance of stock holdings,
since I expect the decision to make a stock trade to represent a stronger opinion than the
decision to hold onto an existing position in a stock. The decision to hold may be driven by
factors unrelated to performance, such as transaction costs or capital gains taxes. Thus, I would
expect trading on private information to be more evident through the examination of trades
rather than holdings. This section addresses this issue by examining the performance of stock
purchases and sales by congressmen.
Table 2, Panel A, displays buy-and-hold returns on various stock portfolios formed based
on aggregate congressmen trades. Specifically, it presents returns on the aggregate portfolio of
all stocks bought by congressmen (Buys), the aggregate portfolio of all stocks sold by
congressmen (Sells) and the hedged portfolio (Buys – Sells). These portfolios are formed each
month based on congressmen trades for that month, according to their financial disclosure
forms. The returns on each portfolio are computed as the buy and hold return that would be
earned through a strategy of purchasing the aggregate congressmen purchases of each stock
during the formation month and selling the aggregate congressmen sells of each stock during
the formation month (month 0). I report one, three, six, and twelve month buy and hold
14

Table 2.2: Performance of Stock Traded by Members of Congress
Panel A: Gross Returns

ALL CONGRESSMEN
Buys
Sells
Buys-Sells
HOUSE
Buys
Sells
Buys-Sells
SENATE
Buys
Sells
Buys-Sells

Month+1 through
Month+6

Month+1 through
Month+12

2.16181**
(2.0871)
2.61588***
(2.7181)
.45407
(0.6996)

4.4421***
(2.9026)
5.34125***
(3.2036)
-.89914
(0.8519)

8.59182***
(4.3506)
10.25542***
(4.4574)
-1.6636
(1.3143)

.44277
( 0.7956)
1.11057**
(2.1239)
-.6678*
(1.7589)

2.08972*
(1.9566)
2.78384***
(2.6620)
-.69412
(0.9491)

4.26807***
(2.6407)
5.08456***
(2.8722)
-.81649
(0.7078)

10.02381***
(5.0699)
9.12845***
(3.6824)
.89536
(0.6380)

.61168
(1.1226)
.52099
(1.2818)
.0907
(0.2090)

2.0947**
(1.9985)
.86738
(1.0522)
1.22732*
(1.8178)

4.36116***
(2.9380)
2.97959**
(2.1543)
1.38156
(1.4368)

6.86986***
(3.3312)
6.95525***
(3.7722)
-.08539
(0.0662)

Month 0

Month +1

Month+1 through
Month+3

Month+1 through
Month+6

Month+1 through
Month+12

-.0193
(0.0786)
.57073**
(2.0872)
-.59003
(1.5669)

-.02793
(0.1644)
.48721**
(2.1411)
-.51514*
(1.8973)

.35155
(1.1950)
.43852
(1.0092)
-.08698
(0.1609)

.26587
(0.5491)
.95883
(1.2320)
-.69297
(0.7625)

-.27475
(0.5449)
1.87259*
(1.7693)
-2.14734*
(1.9153)

.23531
(0.8580)
.41199
(1.3448)
-.17667
(0.4263)

-.06856
(0.2808)
.49665
(1.7759)
-.56521
(1.5175)

.31343
(0.8744)
.43144
(0.8363)
-.11802
(0.1982)

.16662
(0.2988)
.43361
(0.4826)
-.26699
(0.2677)

.93663*
(1.8626)
.4891
(0.4059)
.44753
(0.3627)

-.31862
(1.1757)
.48148*
(1.8701)
-.8001**
(2.4078)

.18432
(0.8770)
.09208
(0.3662)
.09225
(0.2629)

.42981
(1.1615)
-.38469
(1.1298)
.8145
(1.5535)

.3635
(0.6923)
.21134
(0.4207)
.15216
(0.2015)

-1.30111**
(2.0433)
1.17979 *
(1.7859)
-2.4809***
(2.7830)

Month 0

Month +1

.36474
(0.6189)
1.20775**
(2.1657)
-.84301*
(1.8751)

.4048
(0.7657)
1.09182**
(2.2888)
-.68703**
(2.3210)

.6638
(1.1154)
1.03618*
(1.7394)
-.37237
(0.7649)
.09713
(0.1606)
.94501*
(1.7250)
-.84788**
(2.0282)

Month+1 through
Month+3

Panel B: DGTW Adjusted Returns

ALL CONGRESSMEN
Buys
Sells
Buys-Sells
HOUSE
Buys
Sells
Buys-Sells
SENATE
Buys
Sells
Buys-Sells

In Panel A, I compute the buy and hold return on the portfolios of all stocks bought and sold in aggregate by all congressmen.
Buy and hold returns on trade portfolios are based on mimicking the buys and sells by congressmen during each calendar month.
The portfolio formation period is labeled month 0. Panel B presents buy and hold adjusted returns where each buy and hold
return is adjusted by subtracting the buy and hold return on the matching DGTW characteristic portfolio during the holding
period. The reported numbers are the average portfolio gross return in Panel A and the average DGTW adjusted returns in Panel
B during various holding periods following the formation month. The portfolios are weighted based on month 0 shares traded of
each stock. The returns are reported in percent per holding period. *, **, and *** indicate significance at the 10%, 5% and 1%
levels, respectively.

15

returns, averaged across all event months. For example, the average “All Congressmen” return
reported for Month 1(-.687%) is the average monthly buy-and-hold return for a strategy that
mimics the aggregate purchases and aggregate sales of congressmen.
Table 2, Panel B presents benchmark adjusted returns measured with respect to the
portfolio benchmarks developed by Daniel, Grinblatt, Titman, and Wermers ((1997) henceforth
referred to as abbreviation DGTW). To form the DGTW benchmarks, start with all stocks that
have book equity values listed in Compustat and stock returns and market capitalization of
equity listed in CRSP. These stocks are then ranked based on their market capitalization and
assigned to size quintiles (using NYSE size quintile breakpoints). Within each quintile, the stocks
are assigned to book-to-market quintiles creating 25 size and book-to-market fractals. The
stocks in each of these 25 fractals are again sorted into quintiles based on the prior 12-month
return of each stock, resulting in 125 portfolios. For further details on the development of
these portfolios refer to DGTW. These benchmark portfolios are reconstructed at the end of
each June in the manner described above.
The benchmark for each stock is the portfolio to which it belongs. The benchmarkadjusted return for each stock is the difference between the stock return and its benchmark
portfolio return over the same holding period. This return is referred to as the DGTW-adjusted
return.
The results in Table 2, Panel B do not support the hypothesis that congressmen trade on
insider government information, in the aggregate. In fact, the results suggest that Senators
make poor investment decisions. The difference in abnormal returns between buys and sells
16

for month+1 through month+12 for Senators is -2.48%, which is significantly negative at the 1%
level. For the House, the difference between buys and sells for the same holding period was
positive, but not significantly different from zero. This mediocre performance is surprising
given that congressmen are privy to information that the general public is not. However, it is
possibly due to overconfidence, trend chasing, and other common failings of individual
investors (Odean, 1999; Barber and Odean, 2000; Barberis and Thaler, 2003; Barber et al., 2009;
Hoechle et al., 2009)
2.6 Congressmen Net Worth
Finding that congressmen earn negative abnormal returns on their stock transactions is
surprising given the results of Ziabrowski et al.(2004, 2011) and much of the research in political
economy which suggests firms and politicians mutually benefit from their relationships with
each other. However, it is likely that congressmen are taking advantage of their position in
government to increase their income in other ways besides trading on private information in
the stock market. There is substantial evidence that congressmen may in fact reap the benefits
of their positions in Congress in other ways besides investing on private information. For
example, Representative Dennis Hastert purchased land near a proposed highway project and
then inserted a $207 million earmark into the federal highway bill just two months after making
the purchase. The land he purchased for $15,000/acre was sold for $36,000/acre only a year
after he made the purchase.9 Several congressmen have also made large profits through

9

McNamara, M. (2006, June 22). Speaker Hastert’s land deal questioned. CBS News. Retrieved from
http://cbsnews.com

17

gaining access to IPOs, including Rep. John LaFalce, Rep. Gary Ackerman, and Sen. Barbara
Boxer.10
From 2004-2012, the average member of congress saw an inflation adjusted annual
percentage increase in their net worth of 28.9% per year and the median increase of 0.29% per
year.11 The net worth estimates used to calculate the change in net worth of each congressman
were obtained from the Center for Responsive Politics (CRP). The net worth was calculated by
summing each congressman’s assets and subtracting any liabilities reported in their FDRs
(financial disclosure reports). Congressmen report the value of each of their assets,
transactions and liabilities within some range, as described in section II. To determine each
congressman’s net worth, the CRP summed the minimum possible values for each asset, the
maximum possible values for each asset, the minimum liability amount, and the maximum
liability amount. The maximum debt figure was then subtracted from the minimum asset figure
and the minimum debt figure was subtracted from the maximum asset figure. The average of
these two limits is assumed to be their net worth.
There are some shortcomings in disclosure rules that can affect these estimates.
Personal residences that do not produce income are not reported and other personal property,
such as cars or artwork are not reported unless they are owned for investment purposes.
Retirement accounts from employment with the federal government are not reported.
Although these types of assets are not required to be disclosed, some congressmen report
them and when they do report them, these values are included.
10

Whelan, C. B. and Hamburger, T. (2002, September 6). Lawmakers joined executives in profiting from IPO
access. The Wall Street Journal. Retrieved from http://www.wsj.com
11
Since Representative Chellie Pingree’s large net worth increase was due to her marriage, she was removed when
calculating the average annual percentage increase.

18

Table 2.3: House and Senate Net Worth (Top 15 annual % Increase)
Congressman
HOUSE
Pingree, Chellie
Rooney, Tom
Pierluisi, Pedro
Murphy, Patrick
Fallin, Mary
Ney, Bob
Veasey, Marc
Hahn, Janice
Denham, Jeff
Chu, Judy
Gutierrez, Luis V
Ramstad, Jim
Hastert, Dennis
Nussle, Jim
McNerney, Jerry
SENATE
Santorum, Rick
Blunt, Roy
Chambliss, Saxby
Collins, Susan M
Obama, Barack
Feingold, Russ
Lott, Trent
Cruz, Ted
Domenici, Pete V
McConnell, Mitch
Harkin, Tom
Burr, Richard
Jeffords, James M
Dodd, Chris
Allen, George

Year
Elected

End
Year

Increase in dollars
($2012)

Total %
Increase

Annual %
Increase

2008
2008
2008
2012
2006
1994
2012
2011
2010
2009
1992
1990
1986
1990
2006

2012
2012
2012
2012
2010
2006
2012
2012
2012
2012
2012
2008
2006
2007
2012

40500000
11900000
2657073
2979288
3703853
26686
208072
430987
15000000
2114401
2406312
54200000
4680970
748981
365779

365149
31966
6826
1448
5705
2193
994
870
1981
2155
4168
1607
776
1113
2007

73030
6393
1707
1448
1426
1097
994
870
660
539
521
402
388
371
335

1994
2010
2002
1996
2004
1992
1988
2012
1972
1984
1984
2004
1988
1980
2000

2006
2012
2012
2012
2012
2010
2007
2012
2008
2012
2012
2012
2006
2010
2006

1168146
3188959
224057
2635236
4001867
69757
1545954
1398788
3082653
19100000
14800000
2542326
224640
1253243
2114016

694
1177
1152
1103
1097
717
320
82
267
512
461
413
87
251
80

347
147
144
138
137
120
107
82
67
64
58
52
44
42
40

This table displays the increase in net worth, total percentage increase in net worth, and the annual percentage increase in net
worth for the 15 Representatives and Senators who had the largest annual percentage increase in their net worth from 20042012. If a congressman was elected after 2004, I used the data available from the time she takes office, and if a congressman
leaves office prior to 2012, their net worth data ends at the time of their exit. The dollar values used for these calculations are
adjusted for inflation

The averages are adjusted for inflation to observe real changes in net worth. These net worth
estimates include assets unrelated to a member’s gains through Congress, such as assets gained
through marriage, inheritance, or family businesses. Also, since most of the congressmen
19

included in the calculation of the average increase in net worth have been in office longer than
the eight years considered, the real change in net worth for each member during their time in
office may be higher than 28.9% per year.
Table 3 shows the average change in net worth in 2012 dollars, the total percentage
change in net worth, and the annual percentage change in net worth for the 15 Representatives
and Senators whose annual percentage increase in net worth was the highest. The average
Senator saw an increase of 10.8% and the average House member saw an increase of 30.8%.
Each congressman’s average yearly percentage gain was found by dividing their total net worth
growth by the number of years included in the calculation. For example, Representative Patrick
J. Kennedy’s total net worth growth is divided by 6 (2004-2010), since those are the years over
which the growth of his net worth is calculated. Congressmen have large average annual
percentage increases in net worth during their time in office and earn mediocre returns in the
stock market. This suggests that congressmen are earning income in other ways besides
trading on private government information. This is a possible explanation for congressmen’s
mediocre performance in the stock market.
2.7 The Performance of Stocks Traded by Congressmen Close to Retirement
Since the House Committee on Ethics doesn’t have jurisdiction over former members,
departure from the House ends an ethics investigation. Thus, it is relatively safe for House
members to “insider trade” during their final term since, if they are investigated by the ethics
committee, the investigation will end as soon as they leave office. I expect House members to

20

invest on nonpublic information during their last few years in office, resulting in superior
returns on their investments.
Table 4, Panel B reports the difference in DGTW-adjusted returns between aggregate
House member buys and sells 2, 4, 6, and 8 years prior to retirement. The results in Panel A
support the hypotheses that House members trade on private information during their last
term in office. House member earn significant positive abnormal returns on their Buy-Sell
portfolios during the 6 and 12 month holding periods. The abnormal returns on these
portfolios are 4.9% and 8.2%, respectively. The abnormal returns on the buy-sell portfolios 2
years before retirement are also significantly greater than the returns on the buy-sell portfolios
8 years before retirement, further supporting the hypothesis that House members choose to
trade on their private information during their final term. These results also strongly support
evidence in the political agency literature, that term limited politicians change their behavior
(shirk) in their final term (e.g. Besley and Case, 1995).
Unlike the House committee, the Senate committee has no “statute of limitations” for
investigations of past violations. Thus, I suspect Senators are less likely to invest on private
information at any point during their political career. Table 4, Panel B supports this hypothesis.
The difference in abnormal returns between Senator buys and sells are not significantly
different from zero during their final years prior to retirement. This difference in results is likely
due to the fact that the Senate Ethics Committee has no statute of limitations, while the House
Ethics Committee may not begin an ethics investigation once a Representative has left office, or
continue an investigation once the Representative leaves office.

21

Table 2.4: Raw Returns and DGTW adjusted returns 1-8 years before retirement

Panel A: Raw Returns
HOUSE
Years to retirement
(1) 1-2 years (Buys)
(Sells)
Buy-Sell
(2) 3-4 years (Buys)
(Sells)
Buy-Sell
(3) 5-6 years
(Sells)
Buy-Sell
(4) 7-8 years
(Sells)
Buy-Sell
(1) - (4)

SENATE
(1) 1-2 years (Buys)
(Sells)
Buy-Sell
(2) 3-4 years (Buys)
(Sells)
Buy-Sell
(3) 5-6 years
(Sells)
Buy-Sell
(4) 7-8 years
(Sells)
Buy-Sell
(1) - (4)

Month 0

Month +1

Month+1
through
Month+3

0.9625
0.0613
0.9013
(1.18)
0.9625
0.6034
0.3592
(0.41)
0.5015
0.1601
0.3415
(0.42)
-0.0927
0.0370
-.12967
(0.29)
1.0309
(1.22)

1.8030
0.0778
1.7252*
(1.71)
1.8030
1.1552
0.6477
(0.65)
-0.7667
-0.5517
-0.2151
(0.29)
0.1371
0.2017
-.0646
(0.17)
1.7898*
(1.71)

4.6482
1.0087
3.6395*
(1.74)
4.6482
2.9525
1.6957
(1.00)
-0.7673
0.0959
-0.8632
(0.67)
-0.0449
0.3159
-.36083
(0.65)
4.0003*
(1.85)

7.3207
0.5880
6.73272*
(1.81)
7.3207
4.6898
2.6309
(0.98)
0.5273
0.8911
-0.3638
(0.25)
-1.1634
0.2554
-1.4188
(1.50)
8.1515**
(2.14)

11.91787
2.9032
9.0147**
(2.05)
11.9179
7.4637
4.4542
(1.27)
3.7610
2.9531
.80782
(0.37)
1.1285
0.0606
1.06792
(0.91)
7.94673*
(1.74)

0.3068
0.1473
0.1595
( 0.15)
0.7817
0.0741
0.7075
(0.65)
-0.0474
0.1338
-.18116
(0.15)
-0.1427
0.2557
-0.3983
(0.76)
.5578
(0.49)

0.0843
0.7801
-0.6958
(0.85)
0.5042
1.2333
-0.72912
(0.70)
0.9628
-0.3431
1.3058
(1.58)
0.1179
0.3880
-0.2701
(0.66)
-.4257
(0.47)

1.8636
1.5565
0.3072
(0.24)
-0.0978
-0.7583
.66052
(0.34)
1.9829
-0.9339
2.9168**
(2.00)
0.9449
0.4587
0.4861
(0.76)
-.1790
(0.12)

3.1843
3.6540
-0.4696
(0.25)
0.3485
-1.0891
1.4376
(0.45)
2.2535
0.9561
1.2974
(0.64)
2.7227
0.6125
2.1102**
(2.01)
-2.57983
(1.16)

9.0031
10.4326
-1.4296
(0.44)
1.3192
0.4491
0.8701
(0.24)
1.7244
2.8865
-1.16217
(0.50)
2.1951
1.9685
0.2267
(0.15)
-1.65624
(0.46)

22

Month+1
through
Month+6

Month+1
through
Month+12

Table 2.4: (cont’d)
Panel B: DGTW
HOUSE
Years to retirement
(1) 1-2 years (Buys)
(Sells)
Buy-Sell
(2) 3-4 years (Buys)
(Sells)
Buy-Sell
(3) 5-6 years
(Sells)
Buy-Sell
(4) 7-8 years
(Sells)
Buy-Sell
(1) - (4)

Month 0

Month +1

Month+1
through
Month+3

-0.0767
-0.4375
0.3608
(0.52)
0.2811
-0.4375
.7187
(1.31)
0.1840
0.2490
-.6497
(0.09)
-0.2552
-0.2734
.01821
(0.05)
.3426
(0.47)

0.8753
-0.2942
1.169
(1.66)
0.1893
-0.2942
.4834
(0.89)
-0.7087
-0.4091
-.2995
(0.45)
-0.0095
0.0823
-.0919
(0.23)
1.2613*
(1.69)

1.9211
-0.3851
2.3062
(1.50)
0.3458
-0.3851
.7309
(0.89)
-1.0908
-0.2681
-.8227
(0.68)
-0.4215
0.0678
-.4893
(0.84)
2.7955*
(1.74)

SENATE
(1) 1-2 years (Buys)
(Sells)
Buy-Sell

Month+1
through
Month+6

Month+1
through
Month+12

2.8998
-2.0133
4.9132*
(1.74)
-0.0448
-2.0133
1.9685
(1.59)
-0.3251
0.1649
-.4899
(0.36)
-2.0431
-0.3315
-1.7116*
(1.89)
6.6248**
(2.30)

4.64639
-3.5508
8.19718**
(2.39)
-1.2360
-3.5508
2.3148
(1.14)
1.1507
1.1553
-.0046
(0.00)
-0.4132
-0.1512
-.2620
(0.35)
8.4592**
(2.46)

0.6454
-0.2148
-0.1773
-0.7078
-0.1981
1.3993
0.2237
-0.1033
0.7574
2.5933
-0.7539
-.4385
-.0740
-1.4653
-2.7914
(0.81)
(0.62)
(0.07)
(0.91)
(1.22)
(2) 3-4 years (Buys)
0.4340
0.0799
-1.3145
-1.7752
-3.2388
(Sells)
0.1095
1.0073
-1.5566
-3.2538
-2.1590
Buy-Sell
.3246
-.9274
.2421
1.4785
-1.0798
(0.33)
(0.95)
(0.15)
(0.52)
(0.35)
(3) 5-6 years (Buys)
-0.3347
0.7016
0.7008
0.4221
-2.2383
(Sells)
0.3790
-0.6759
-1.6802
-1.0912
-0.7809
Buy-Sell
-.7137
1.3775*
2.3810*
1.5133
-1.4574
(0.67)
(1.90)
(1.87)
(0.77)
(0.59)
(4) 7-8 years (Buys)
-0.4686
-0.0805
0.1205
1.2673
0.0570
(Sells)
0.1269
0.0744
-0.3006
-1.1418
-0.8210
-.5955
-.1548
.4211
2.4091**
.8779
Buy-Sell
(1.13)
(0.39)
(0.69)
(2.39)
(0.60)
(1) - (4)
-.1584
-.2837
-.4951
-3.8744**
-3.6693
(0.15)
(0.37)
(0.44)
(2.06)
(1.31)
The reported numbers are the average DGTW adjusted returns during various holding periods following the
formation month. The portfolios are weighted based on month 0 shares traded of each stock. Panel A reports
returns for House Members 2, 4, 6, and 8 years prior to retirement. It also reports the difference between the
buy-sell portfolios 2 years prior to retirement and the buy-sell portfolios 8 years prior to retirement. Panel B
reports similar results for Senate members. The returns are reported in percent per holding period. *, **, ***
indicate significance at the 10%, 5% and 1% levels, respectively.

23

2.8 Alternative Explanations for Abnormal Returns
Previous studies have analyzed common stock returns for several subsamples of
congressmen. Ziabrowski et al. (2004) find there are no significant differences between the
abnormal returns of Democrats and Republicans, but seniority is a significant factor. The
common stock investments of Senators with the least seniority (serving less than seven years)
significantly outperform the investments of the most senior Senators (serving more than 16
years). Eggers and Hainmueller (2014) investigate whether political relationships between
members and firms impact politician portfolio choices. They find that congressmen stock
holdings are skewed toward firms in their home districts and firms whose PACs gave them
campaign contributions. However, they do not disproportionately invest in companies to which
they are connected through their committee assignments. Congressmen’s returns on their
investments in firms in which they have a connection (firms located in a congressman’s home
district or that made campaign contributions) tend to outperform their returns on investments
in firms with no connection.
I also analyze common stock returns for other subsamples. Specifically I analyze
whether abnormal returns vary for Senate and House members depending on their net worth
and seniority.
2.8.1 Net Worth

I expect that a congressman with a lower net worth will feel like she has more to gain
from investing on private information than a congressman with a high net worth. Table 5
describes the distribution of estimated net worths across all congressmen, House members,
and Senators. Each year, I assign the congressmen’s net worths to quintiles, where quintile 1
24

(Q1) is the lowest net worth and quintile 5 (Q5) is the highest. This is done for all congressmen
and separately for House members and Senators. Not surprisingly, the Senators tend to be
wealthier than the Representatives. If net worth impacts whether congressmen trade on their
private information then I expect the effect to be stronger in the House.

Table 2.5: Summary of Congressmen Net Worth
Mean

Min

Max

Median

ALL CONGRESS
Q1
Q2
Q3
Q4
Q5

-50313.77
373755.6
924998.5
2647182
31100000

-19000000
165505
514511
1181301
3886521

236013
647015
1695514
4694011
501000000

60503
374003
870508
2515658
10500000

HOUSE
Q1
Q2
Q3
Q4
Q5

-93449.34
312478
770993.7
2158189
24,000,000

-19000000
115502
430504
991524
3257041

199506
525012
1444009
4005007
5.01E+08

42361
306503
745757.5
2028011
7751982

SENATE
Q1
Q2
Q3
Q4
Q5

236665.7
834578.4
2211478
5464595
61700000

-1767771
466015
998003
2405516
7881552

576008
1640031
3740014
12100000
337000000

247503
773506
2000528
5182015
28500000

Each year, congressmen’s estimated net worth is divided into quintiles, where Q1 is the lowest net worth and Q5 is the highest.
This is done for all congressmen, the House members, and Senators. Net worth estimates are from the Center for Responsive
Politics.

Table 6 reports the raw and DGTW-adjusted returns for the aggregate portfolio of all
stocks bought (buys) or sold (sells) by House members or Senators in each net worth quintile,
and the hedged portfolio (Buys – Sells). The most noteworthy results are for the House
member net worths that fall in Quintile 3. Specifically, House members with an average net
worth of $770,994 earned significantly positive DGTW adjusted returns, at the 5% level, on

25

Table 2.6: Raw and DGTW Adjusted Returns for Net Worth Quintiles
Panel A:
Raw Returns
HOUSE
Q1
Buys
Sells
Buy-Sell
Q2
Buys
Sells
Buy-Sell
Q3
Buys
Sells
Buy-Sell
Q4
Buys
Sells
Buy-Sell
Q5
Buys
Sells
Buy-Sell
Q1-Q5

Month 0

Month +1

Month+1 through
Month+3

Month+1 through
Month+6

Month+1 through
Month+12

0.1370
1.4888
-1.35174
(1.62)

0.7882
1.5626
-0.7745
(0.93)

2.6007
2.9924
-0.3917
(0.24)

4.9407
4.4268
0.5139
(0.25)

9.9261
11.9391
-2.0129
(0.58)

1.1442
0.3705
0.7737
(0.65)

-0.4093
-0.3801
-0.0292
(0.04)

-1.7462
0.5073
-2.2535
(1.42)

-0.9153
2.6871
-3.6024*
(1.73)

1.7071
4.4715
-2.76438
(0.95)

0.3424
2.2008
-1.8585*
(1.81)

0.3209
1.0741
-0.7532
(1.07)

3.0498
0.9231
2.1266*
(1.75)

5.9011
2.5922
3.3089*
(1.69)

11.6619
5.9098
5.7521*
(1.90)

-0.1930
0.1454
-0.3384
(0.64)

0.9179
0.5829
0.3350
(0.59)

2.7936
1.1095
1.6841*
(1.66)

6.0897
3.5669
2.5228*
(1.66)

9.8057
7.0975
2.7082*
(1.69)

0.7663
0.9712
-0.2048
(0.40)

0.3188
1.1328
-.8140*
(1.92)

1.9414
2.9502
-1.009
(1.27)

4.1818
5.2922
-1.1104
(0.92)

10.2208
9.2799
0.9408
(0.62)

-1.1469
(1.20)

.03947
(0.04)

.61707
(0.33)

1.62432
(0.73)

-2.95376
(0.83)

0.8999
-1.0493
1.9492**
(2.33)

1.1403
-0.7704
1.9107**
(2.23)

1.1538
-0.6307
1.7845*
(1.68)

1.2883
-1.3169
2.6052*
(1.88)

0.4992
0.7555
-0.2564
(0.12)

0.55254
0.87734
-0.3248
(0.36)

-0.1018
0.6141
-.71586
(1.07)

0.0022
-0.2881
0.2903
(0.26)

0.4247
0.7103
-0.2856
(0.18)

2.0374
2.5666
-0.5293
(0.20)

1.8863
0.3089
1.5774*
(1.81)

0.98055
0.42345
0.55711
(0.66)

2.149
2.0494
0.0996
(0.08)

4.8928
3.8923
1.0005
(0.57)

7.7807
6.3913
1.3894
(0.54)

1.3449
0.4503
0.8946
(0.95)

0.9930
-0.0581
1.0511
(1.62)

3.0193
-0.5675
3.5868***
(3.04)

5.2492
0.6157
4.6334***
(3.57)

7.6734
3.7520
3.9214*
(1.96)

-0.1969
1.1363
-1.3332***
(2.73)

0.6920
0.7834
-0.0914
(0.20)

2.4017
1.2960
1.1058
(1.49)

4.4709
3.2007
1.2703
(1.15)

6.7799
6.4234
.3565
(0.24)

3.2824***
(3.42)

2.0022**
(2.10)

.6787
(0.54)

1.3349
(0.77)

-.61291
(0.23)

SENATE
Q1
Buys
Sells
Buy-Sell
Q2
Buys
Sells
Buy-Sell
Q3
Buys
Sells
Buy-Sell
Q4
Buys
Sells
Buy-Sell
Q5
Buys
Sells
Buy-Sell
Q1-Q5

26

Table 2.6: (cont’d)
Panel B: DGTW
HOUSE
Q1
Buys
Sells
Buy-Sell
Q2
Buys
Sells
Buy-Sell
Q3
Buys
Sells
Buy-Sell
Q4
Buys
Sells
Buy-Sell
Q5
Buys
Sells
Buy-Sell
Q1-Q5

SENATE
Q1
Buys
Sells
Buy-Sell
Q2
Buys
Sells
Buy-Sell
Q3
Buys
Sells
Buy-Sell
Q4
Buys
Sells
Buy-Sell
Q5
Buys
Sells
Buy-Sell

Month 0

Month +1

Month+1 through
Month+3

Month+1 through
Month+6

Month+1 through
Month+12

-0.5721
0.8279
-1.4001**
(2.31)

0.3238
0.8829
-0.5591
(0.83)

0.3544
1.3707
-1.0163
(0.84)

0.5325
1.1507
-0.6182
(0.36)

0.7019
4.3410
-3.6390
(1.32)

0.8463
-0.1282
0.9745
(0.86)

-0.7916
-0.4738
-0.3178
(0.52)

-2.6655
-0.8091
-1.8565
(1.22)

-3.9344
-0.9896
-2.9448
(1.26)

-5.9121
-0.7597
-5.1524*
(1.89)

0.2533
1.6433
-1.3901
(1.49)

0.1305
0.6585
-0.5280
(0.87)

1.8229
-0.5571
2.3800**
(2.28)

2.6735
-1.4109
4.0844**
(2.43)

4.1972
-2.4864
6.6836**
(2.51)

-0.5229
-0.1192
-0.4037
(0.79)

0.1729
0.0410
0.1319
(0.28)

0.4797
-0.8305
1.3102
( 1.50)

1.6875
-0.5272
2.2147*
( 1.72)

0.8988
-1.6847
2.5834*
(1.81)

0.31985
0.34151
-0.02166
(0.05)

-0.1717
0.5088
-0.6805*
(1.66)

0.2124
0.5471
-0.3347
(0.51)

0.0740
0.5125
-0.4386
(0.42)

1.03113
0.52477
0.50636
(0.38)

-1.3784*
(1.90)

.1214
(0.16)

-.6816
(0.47)

-.1796
(0.10)

-4.1454
(1.47)

0.3694
-0.8285
1.1979*
(1.84)

0.6685
-1.0221
1.6905**
(2.25)

-0.0153
-0.9785
0.9633
(1.0)

-0.4275
-1.7466
1.3191
(0.97)

-1.7033
-0.1848
-1.5185
(0.85)

0.19365
0.39711
-0.20346
(0.27)

-0.8793
0.1223
-1.0016*
(1.75)

-1.3364
-1.5127
0.1763
(0.19)

-2.5341
-2.2030
-0.3311
(0.26)

-3.302
-3.311
0.009
(0.00)

1.0862
0.1207
0.9656
(1.29)

0.25178
-0.18339
0.43518
(0.61)

0.16086
0.03952
0.12134
(0.10)

0.1926
-0.56996
0.76256
(0.46)

-1.8613
-0.1158
-1.7455
(0.75)

0.6049
0.6150
-0.0101
(0.01)

0.1939
-0.4710
0.66487
(1.23)

0.8073
-1.5129
2.3202**
(2.51)

1.2099
-1.2822
2.4921**
(2.16)

-1.6216
-0.4424
-1.1793
(0.68)

-0.5543
0.5593
-1.1137***
(2.94)

0.3088
0.3469
-0.0381
(0.11)

0.7957
0.1296
0.6661
(1.26)

0.5967
0.9295
-0.3328
(0.41)

-1.0761
2.0146
-3.0908***
(3.27)

Q1-Q5
2.3115***
1.7286**
.2972
1.65189
1.57229
(3.17)
(2.05)
(0.26)
(1.03)
(0.77)
The reported numbers are the average Raw (Panel A) and DGTW adjusted (Panel B) returns during various holding periods
following the formation month for each net worth quintile for all congressmen, House members, and Senators. The quintiles are
formed each year and then averaged across all years (2004-2012). The returns are reported in percent per holding period. *, **,
*** indicate significance at the 10%, 5% and 1% levels, respectively.

27

their portfolios during the 3, 6, and 12 month holding periods. The abnormal returns on these
portfolios are 2.37%, 4.08% and 6.68% respectively. House members with net worths that fall
in Quintile 4 also earn smaller, but significantly positive abnormal returns during the 6 and 12
month holding periods of 2.21% and 2.58%, respectively. The Senators who have a net worth in
Quintile 4 earn significantly positive abnormal returns, at the 5% level, on their buy-sell
portfolios during the 3 and 6 month holding periods of 2.32% and 2.49%, respectively.
However, Senators with a net worth in Quintile 5, the richest Senators, earned a
significantly negative return at the 1% level of -3.09% for the 12 month holding period. These
results suggest that congressman, and especially the House members, with a net worth in the
fourth quintile make better investment decisions than the congressman in other quintiles.
2.8.2 Seniority
I expect the most senior congressmen to make more informed investments than those
congressmen with the least seniority. Positions of power within Congress, such as committee
memberships, are generally determined by party and seniority. Also, it is likely that
congressmen develop more and better connections that will gain them access to insider
information as they increase their experience in Congress.
Table 7, Panel A reports the DGTW-adjusted returns for the aggregate portfolio of all
stocks bought (buys) or sold (sells) by House members 2 through 16 years, and greater than 16
years after being voted into office. The Table also reports the hedged portfolios (Buys – Sells).
Panel B reports the same results for Senators. The most surprising results are in the House.
House members earn significant positive abnormal returns, 1 to 2 years after being voted into
28

Table 2.7: DGTW Adjusted Returns for Various Seniorities

Month 0
Panel A: HOUSE
Years following initial election
(1) 1-2 years (Buys)
-0.0465
Sells
0.6332
Buy-Sell
-0.6797
(2) 3-4 years (Buys)
Sells
Buy-Sell
(3) 5-6 years (Buys)
Sells
Buy-Sell
(4) 7-8 years (Buys)
Sells
Buy-Sell
(5) 9-10 years (Buys)
Sells
Buy-Sell
(6) 11-12 years (Buys)
Sells
Buy-Sell
(7) 13-14 years (Buys)
Sells
Buy-Sell
(8) 15-16 years (Buys)
Sells
Buy-Sell
(9) >16 Terms (Buys)
Sells
Buy-Sell
(1) - (9)

(0.97)
0.0370
-0.4047
0.4417
(1.05)
0.97008
-0.57052
1.5406***
(2.81)
-0.2481
-0.0354
-0.2127
(0.35)
-0.5410
-0.3059
-0.2351
(0.36)
-0.1346
-0.1798
0.0452
(0.07)
-0.0401
0.3928
-0.4329
(0.66)
-0.9015
0.3843
-1.2859
(2.77)
1.2419
-0.0106
1.2525
(1.64)
-1.9322*
(1.71)

Month +1

Month+1
through
Month+3

Month+1
through
Month+6

Month+1
through
Month+12

0.0078
0.7499
-0.7421
(1.05)
0.0284
0.1733
-0.1448
(0.43)
0.58023
-0.1812
0.76143
(1.61)
-0.52629
-0.12297
-0.40331
(0.80)
-0.2190
0.8026
-1.0216
(1.23)
0.2152
0.0656
0.1495
(0.29)
-0.3561
-0.7539
0.3978
(0.74)
-0.5506
0.0826
-0.6332
(1.03)
0.1598
0.2071
-0.0474
(0.10)
-0.6947
(0.77)

0.1104
-0.5097
0.6201
(0.53)
0.5547
-0.0447
0.5994
(1.00)
0.0976
-0.2774
0.3750
(0.41)
-0.3831
1.5495
-1.9326
(1.62)
-0.2261
1.1245
-1.3505
(1.05)
0.7208
-1.0831
1.8038
(1.79)
-0.2416
-0.8712
0.6296
(0.61)
0.1088
0.1169
-0.0081
(0.01)
-0.1548
0.8647
-1.0194
(1.10)
1.6395
(0.95)

0.9036
-2.8644
3.768*
(1.83)
0.4546
0.1303
0.3243
(0.35)
0.5297
-1.3728
1.9025
(1.47)
-0.8381
0.3652
-1.2033
(1.00)
-1.8535
1.2023
-3.0558*
(1.91)
0.6935
-0.9838
1.6773
(1.03)
-0.0825
-1.5688
1.4864
(1.03)
0.2894
0.5263
-0.2369
(0.13)
-1.5413
1.4785
-3.0198**
(2.26)
6.7878**
(2.49)

0.0408
-3.4844
3.5252**
(1.99)
0.4535
0.1561
0.2974
(0.27)
1.5784
-1.3533
2.9316*
(1.86)
-1.9374
0.8797
-2.8171
(1.61)
0.7271
1.3755
-0.6484
(0.38)
0.274
-1.0918
1.3658
(0.58)
0.2512
-2.9801
3.2313
( 1.59)
0.9431
0.3008
0.6423
(0.22)
-0.8370
2.3621
-3.1991
(1.64)
6.7244**
(2.40)

29

Table 2.7: (cont’d)

Month 0
Panel B: SENATE
Years following initial election
(1) 1-2 years (Buys)
0.3824
Sells
-0.7833
Buy-Sell
1.1657*
(1.81)
(2) 3-4 years (Buys)
0.2954
Sells
0.5675
Buy-Sell
-0.2722
(0.33)
(3) 5-6 years (Buys)
0.5570
Sells
0.0920
Buy-Sell
0.4650
(0.50)
(4) 7-8 years (Buys)
0.4331
Sells
0.2018
Buy-Sell
0.2312
(0.27)
(5) 9-10 years (Buys)
0.3866
Sells
0.0728
Buy-Sell
0.3138
(0.49)
(6) 11-12 years (Buys)
0.4029
Sells
0.0109
Buy-Sell
0.3920
(0.42)
(7) 13-14 years (Buys)
-0.2885
Sells
0.0434
Buy-Sell
-0.3319
(0.63)
(8) 15-16 years (Buys)
-0.6008
Sells
-0.7089
Buy-Sell
0.1081
(0.16)
(9) >16 Terms (Buys)
-0.3825
Sells
0.4784
Buy-Sell
-0.8609**
(2.01)
(1) - (9)
2.0266**
(2.49)

Month +1

Month+1
through
Month+3

Month+1
through
Month+6

Month+1
through
Month+12

-0.3751
-0.3205
-0.0546
(0.09)
0.5037
0.4587
0.0450
(0.08)
-0.1619
0.0419
-0.2038
(0.33)
0.4200
0.6000
-0.1801
(0.26)
1.0367
0.5455
0.4912
(0.54)
-0.6480
-0.7232
0.0752
(0.12)
0.4624
-0.8608
1.3232
(1.33)
0.2110
-0.0872
0.2982
(0.59)
0.3627
0.2832
0.0795
(0.20)
-0.1340
(0.19)

-0.2635
-0.1383
-0.1253
(0.13)
-0.0418
-0.0690
0.0272
(0.03)
-0.2539
0.6985
-0.9525
(0.86)
-1.0315
-0.0914
-0.9400
(0.91)
0.0707
-2.3646
2.4353*
(1.97)
-0.3923
-0.6384
0.2461
(0.19)
-0.0869
-1.4275
1.3406
(1.19)
0.9972
0.2107
0.7865
(1.03)
0.6610
0.0703
0.5907
(1.14)
-0.7159
(0.64)

-0.5518
-0.4983
-.05359
(0.04)
-0.3519
0.5512
-0.9031
(0.63)
-0.6321
-0.3607
-0.2714
(0.17)
-0.6464
-0.4808
-0.1656
(0.12)
0.2839
-3.1489
3.4328*
(1.83)
-1.3086
-0.9841
-0.3246
(0.20)
-1.0451
-1.5753
0.5303
(0.34)
1.6892
1.8298
-0.1405
(0.10)
0.6715
0.7278
-0.0564
(0.06)
.0028
(0.00)

0.3375
1.5007
-1.1632
(0.51)
-1.6619
-1.6862
0.0243
(0.01)
-0.7701
1.2118
-1.9819
(0.95)
-0.1319
-1.0555
0.9237
(0.47)
-1.1535
-1.1386
-0.0150
(0.01)
-3.7256
-2.3519
-1.3737
(0.78)
-1.7937
-1.4603
-0.3335
(0.17)
1.4466
1.4801
-0.0335
(0.02)
-1.2064
2.5141
-3.7205***
(3.51)
2.5573
(0.92)

The reported numbers are the average DGTW adjusted returns during various holding periods following the formation month.
The portfolios are weighted based on month 0 shares traded of each stock. Panel A reports returns for House Members 2, 4, 6, 8,
10, 12, 14, and 16 years after being voted into office. It also reports the difference between the buy-sell portfolio returns 2 years
after being voted into office and the buy-sell portfolio returns greater than 16 years after being voted into office. Panel B reports
similar results for Senate members. The returns are reported in percent per holding period. *, **, *** indicate significance at the
10%, 5% and 1% levels, respectively.

30

office, on their hedged portfolios during the 6 and 12 month holding periods of 3.77% and
3.53%, respectively. There is also a significant difference between the abnormal returns earned
2 years following being elected in the House and more than 16 years after being elected for the
6 and 12 month holding periods. The differences are 6.79% for the 6-month holding period and
6.72% for the 12-month holding period, which are both significant at the 5% level.
These results are surprising for a couple of reasons. First, I expected the more
experienced House members to earn better returns on their common stock transactions than
the less experienced House members. Second, the most experienced House members do not
earn abnormal returns, but the earlier results suggest House members perform well during
their final years prior to retirement. It is important to note that only the House members who
choose to retire perform well during their final term. However, the House members who leave
office unwillingly because they failed to get reelected do not perform well during their final
term, on average.
2.9 Conclusion
In this chapter, I examine the hypothesis that politicians try to maintain a good
reputation while they care about reelection, but act corruptly once they decide to retire, thus
imposing a term limit. Using a dataset on congressmen’s common stock transactions, I find that
House members earn abnormal returns of 4.9% and 8.2% for a 6 and 12-month holding period,
respectively, during their final term in office. However, Senators do not earn abnormal returns
during the 2, 4, 6, or 8 years prior to their retirement. The difference in results between the

31

House and Senate is likely caused by the difference in the laws governing the House and the
Senate.
I also examine whether abnormal returns vary for Senate and House members
depending on their net worth and seniority. The most notable result is that House members
earn abnormal returns of 3.5% for the 12-month holding period during their first term in office.
Otherwise, House members and Senators earn mediocre or negative returns.
This chapter provides evidence that trading on private political information should be
regulated. I provide evidence that House members trade on private political information during
their final term prior to retiring from office, while Senators do not. A policy implication of this
study is that the House Committee on Ethics should have no “statute of limitations” for
investigations of past violations of House members. This may discourage unethical behavior
during the final term in office.

32

CHAPTER 3: Wealth Accumulation by U.S. Congress Members
3.1 Introduction
Do members of the United State Congress enhance their wealth by making investments
based on private political information? There is substantial evidence in support of the
hypothesis that they do. These investments are not only in public corporations, but also in
land, private equity, and/or private business ventures. A myriad of examples of congressmen
making gains on private government information is provided in a widely cited book by Peter
Schweizer (2008) as well as a substantial number of news articles. For example, in 2002 and
2004, House Speaker Dennis Hastert bought farm land in Plano, Illinois before inserting a $207
million earmark12 into the federal highway bill to build major roads near the farm land which
would increase the value of the land as a potential residential area.13 As a result, the value of
the land increased to $36,000 an acre, a 140% profit. This was not considered illegal due to the
fact that Hastert was able to demonstrate that at least one other person would benefit from
the earmark.
A substantial body of research in political science, finance, and economics supports the
idea that politicians may get rich by investing on their private government information. Studies
that focus on countries outside of the U.S. have supported this hypothesis primarily through

12

An earmark is a provision in congressional legislation that allocates a specified amount of money to a specific
project, program, or organization.
13
Another example is Ken Calvert’s earmark for $1.5 million to support commercial development around the
March Air Reserve Base in Southern California, shortly after he and a partner purchased a 4.3-acre parcel of land
just south of the base for $550,000. Less than a year after the earmark, Calvert and his partner sold the land,
without making improvements to it, for $985,000. (Hamburger, T., Pugmire, L., and Simon, R. (2006, May 15). Rep.
Calvert’s Land of Plenty. Los Angeles Times. Retrieved from http://articles.latimes.com)

33

focusing on relationships between politicians and public corporations. For example, it has been
found that political connections help firms gain access to bank loans (Claessens, Feijen, and
Laeven, 2008; Faccio, 2006; Khwaja and Mian, 2005), obtain special tax exemptions (Faccio,
2006), and increase the likelihood of a government bailout (Faccio, Masulis, and McConnell,
2006). Given these findings, it is not surprising that political connections increase firm value
(Fisman, 2001), or improve firm performance (Johnson and Mitton, 2003).
Similar to studies on corruption in foreign countries, the hypothesis that U.S.
congressmen may get rich by investing on their private information has been supported
through focusing on congressional public equity ownership and relationships between U.S.
congressmen and public corporations. It has been found that political connections in the U.S.
help firms receive more government contracts (Tahoun, 2014) and receive help in dealing with
regulatory agencies (Correia, 2009; Yu and Yu, 2006). 14 In addition, Tahoun and van Lent (2010)
found that US congressmen’s equity ownership is positively correlated with their voting in favor
of legislative proposals to bail out the financial sector during the recent financial crisis. Thus,
similar to studies on other countries, it is not surprising that shifts in the controlling party of the
U.S. government have a large impact on the market value of firms dependent on the firm’s
political leaning (Jayachandran, 2006), S&P 500 companies earn positive abnormal stock returns
following the announcement of the nomination of a politically connected individual to the
board (Goldman, Rocholl, and So, 2009), and congressmen’s connected investments generally
outperform their other investments (Eggers and Hainmueller, 2014).

14

Yu and Yu (2006) find that firms’ lobbying activities make a significant difference in fraud detection. Correia
(2009) finds that firms may use political expenditure to avoid investigation and prosecution by the SEC for
misreporting.

34

In this chapter, I investigate whether congressmen act corruptly by focusing on returns
to their total wealth rather than only focusing on their public equity investments. The net
wealth data is obtained from congressmen’s Financial Disclosure Reports, in which they are
required to annually publicly disclose detailed information on their holdings and transactions in
income-producing property and assets such as stocks, bonds, mutual funds, hedge funds,
private equity funds, and real property. This data was obtained from the Center for Responsive
Politics for years 2004 to 2012. The use of total wealth should provide a better estimate of the
extent to which congressmen use private government information for individual gain, since
total wealth measures capture gains to real estate/ land ownership, private equity investments,
and small business ownership. While the actions of public corporations are very much in the
public eye, little is known about the activities of private companies and real estate deals. Thus,
congressmen may be more likely to act corruptly through these outlets due to their private
nature. This idea is supported by Tahoun’s (2014) finding that the positive relationship
between congressional stock ownership and corporate campaign contributions is stronger for
small firms, suggesting that congressmen have a greater ability to provide benefits to firms that
are relatively small since this is less likely to attract public scrutiny.
Several studies examine the wealth accumulation of politicians. Eggers and Hainmueller
(2009) find that Britain’s Conservative Members of Parliament (MPs) make financial gains in
public office while Labour MPs do not. Fisman, Schulz, and Vig (2012) and Bhavnani (2012)
examine the wealth accumulation of Indian parliamentarians using financial disclosures and
find that election winners have a greater rate of wealth accumulation than losers. To the best
of my knowledge only two studies have examined wealth accumulation of U.S. congressmen.
35

Querubin and Snyder (2009) examine the wealth accumulation of U.S. congressmen during
1850-1880 using a regression discontinuity design and find that those who won elections
earned more than those who lost only during 1870-1880. Lenz and Lim (2009) compare the
wealth accumulation of U.S. House members to the wealth accumulation of a matched sample
of non-politicians from the Panel Study on Income Dynamics and find that U.S. representatives
accumulate wealth at the same rate as similar households.15
My study differs from previous examinations of the wealth accumulation of U.S.
congressmen in some key aspects. For example, I use the capital asset pricing model (CAPM)
and Fama-French Three-Factor model to measure whether U.S. congressmen earn abnormal
returns to their wealth. I also use the human capital CAPM (HCAPM) model developed by
Jagannathan and Wang (1996), which significantly improves upon the traditional CAPM. In
addition to measuring congressmen’s abnormal returns to total wealth, I determine whether
these returns are due to market timing or asset selection abilities. That is, can congressmen
successfully predict future market movements to appropriately shift their assets in or out of the
market, and/or can they consistently select undervalued securities? I use two models to
separate market timing and asset selection ability. The first, Treynor and Mazuy’s (henceforth
TM) (1996) model, adds a quadratic term to the traditional CAPM model to measure market
timing skills. The second approach, Henriksson and Merton’s (henceforth HM) (1981) model,
adds a bull and bear market condition to the traditional CAPM to measure market timing skills.

15

The Panel Study of Income Dynamics (PSID) interviewed a nationally representative sample of approximately
3,000 households, to ask for the same financial disclosure information that congressmen are required to disclose.

36

The first main result of this chapter is that congressmen earn large abnormal returns to
their wealth, thus confirming the public perception that congressmen unethically trade on
private government information. Using the CAPM and Fama-French three-factor models I find
that for Congress as a whole, members on average earn abnormal returns to their wealth of
6.68% per year and 6.39% per year, respectively. These results hold when I include a measure
of the aggregate return to human capital to improve upon the traditional CAPM, as suggested
by Jagannathan and Wang (1996). Using the HCAPM model I find that congressmen earn large
abnormal returns to wealth of 8.25%. Similarly, the Fama-French three-factor model with the
inclusion of human capital suggests that congressmen earn abnormal returns to wealth of
7.99%.
The second main result is that congressmen earn abnormal returns due to their asset
selection abilities. Using the Treynor-Mazuy and Henriksson-Merton models, I find that the
average alpha, the measure for selectivity, is 5.93% and 5.53%, respectively. These results
withstand the incorporation of the human capital factor in the TM and HM models. Given
previous research on relationships between congressmen and firms that suggests congressmen
make legislative decisions based on their portfolio holdings, that they invest on their private
government information, and that they earn greater returns on investments in firms from
which they receive campaign contributions, the result that congressmen earn positive abnormal
returns due to their selection abilities is not surprising. 16

16

For examples see Correia (2009), Eggers and Hainmueller (2014), Goldman, Rocholl, and So (2009), Jayachandran
(2006), Tahoun (2014), Tahoun and van Lent (2010), Yu and Yu (2006).

37

I conclude my data analysis with an investigation of whether powerful congressmen
drive the main results that congressmen earn large abnormal returns to their wealth. The
political science literature suggests that members of powerful committees have the ability to
award benefits to their connections (Edwards and Stewart, 2006) and affect the distributions of
government benefits (Roberts, 1990). Thus, members of powerful committees may earn larger
returns to wealth than non-committee members, due to greater access to private information
and greater abilities to bestow benefits to their connections. To rule out the possibility that
powerful politicians are driving the main results of this chapter, I separately measure the
returns to wealth for powerful congressmen using the Fama-French three-factor model with
and without the human capital factor. The estimated returns to wealth for powerful
congressmen suggest that these do not drive the main results.
3.2 Data Description
According to the “Ethics in Government Act of 1978”, members of Congress are required
to file annual reports disclosing detailed information on their financial holdings and
transactions in income-producing property and assets, such as stocks, bonds, mutual funds,
hedge funds, private equity funds, and real property. Personal financial disclosure forms are
filed annually by May 15, covering the preceding calendar year for government officials, their
spouses, and dependent children. Although the financial disclosure reports (FDRs) are not
audited, failure to honestly disclose is a felony and members have been prosecuted. 17
I obtained data on congressmen’s net wealth and financial assets from the annual
financial disclosure reports submitted between 2004 and 2012 by members of Congress and
17

The most recent is Senator Ted Stevens of Alaska.

38

transcribed by the Center for Responsive Politics (CRP)18. The net worth was calculated by
summing each congressman’s assets and subtracting any liabilities reported in their FDRs.
Congressmen report the value of each of their assets and liabilities within broad ranges of $1 to
$1,000, $1,001 to $15,000, $15,001 to $50,000, $50,001 to $100,000, $100,001 to $250,000,
$250,001 to $500,000, $500,001 to $1,000,000, $1,000,001 to $5,000,000, $5,000,000 to
$25,000,000, $25,000,001 to $50,000,000, and over $50,000,000. To determine each
congressman’s net wealth, the CRP summed the minimum possible values for each asset, the
maximum possible values for each asset, the minimum liability amount, and the maximum
liability amount. The maximum debt figure was then subtracted from the minimum asset figure
and the minimum debt figure was subtracted from the maximum asset figure. The average of
these two limits is assumed to be their net worth.
There are some shortcomings in disclosure rules that can affect these estimates.
Personal residences that do not produce income are not reported and other personal property,
such as cars or artwork are not reported unless they are owned for investment purposes.
Retirement accounts from employment with the federal government are not reported.
Although these types of assets are not required to be disclosed, some congressmen report
them and when they do, these values are included.
Overall, my dataset includes 190,684 reported end-of-year assets holdings and
estimated net worth for 1,026 members of Congress. Table 1 provides summary statistics for
net wealth and asset ownership. The asset ownership is broken down into subcategories,

18

I start in 2004 because Financial Disclosure Reports filed since 2004 have been made readily available at
opensecrets.org.

39

Table 3.1: Summary of Net Worth and Assets
Panel A: All Congressmen

2004
2005
2006
2007
2008
2009
2010
2011
2012

Mean Net
Worth
$6,180,330
6,467,371
6,823,783
7,803,586
6,467,679
6,595,942
7,395,009
7,982,701
7,605,058

Median Net
Worth
$743,508
780,014
832,760
834,256
803,507
907,260
967,519
966,002
1,022,762

Real
Estate
10.96%
10.86
12.42
12.82
14.05
17.39
16.43
17.33
18.95

Risky
14.25%
10.17
11.31
15.35
20.19
18.93
21.22
18.68
12.86

Public
Equity
50.41%
56.26
52.86
47.32
32.88
41.02
42.04
43.39
49.00

Safe
17.87%
19.05
19.46
21.80
30.20
21.93
18.75
19.05
18.01

Blind
Trust
6.49%
3.65
3.94
2.71
2.67
0.73
1.57
1.55
1.19

$653,008
656,756
742,155
732,509
698,004
789,012
809,511
858,005
878,509

14.48%
15.32
15.69
15.78
17.59
19.78
18.68
18.35
19.81

12.88%
8.41
12.49
13.95
20.76
21.06
23.44
20.93
12.18

60.03%
64.18
56.69
54.21
33.82
42.70
42.02
43.88
51.77

11.45%
10.97
14.16
15.79
27.78
16.44
14.43
15.44
14.68

1.17%
1.11
0.97
0.27
0.04
0.02
1.42
1.41
1.56

$1,638,341
1,744,024
1,713,044
2,000,528
1,927,285
2,108,781
2,363,023
2,778,529
2,877,550

8.28%
8.66
8.65
8.40
10.98
15.82
13.80
18.24
17.64

19.64%
16.88
13.11
22.00
25.32
18.31
18.71
17.04
15.47

37.33%
38.42
47.01
38.15
30.26
33.12
35.82
37.47
41.02

22.44%
30.44
24.78
27.80
30.03
30.25
29.28
24.74
25.49

12.31%
5.61
6.45
3.65
3.41
2.50
2.40
2.52
0.38

Panel B: House Members
2004
2005
2006
2007
2008
2009
2010
2011
2012

$4,409,041
4,668,715
5,293,304
5,836,054
4,847,519
5,181,048
6,081,902
6,601,313
6,016,726

Panel C: Senators
2004
2005
2006
2007
2008
2009
2010
2011
2012

$13,202,844
13,677,434
13,084,792
16,479,072
13,135,091
12,780,328
13,035,817
14,503,391
14,753,942

Estimates are not adjusted for inflation. Real estate assets include land, residential and commercial buildings, and real estate
partnerships. Risky assets include private businesses, partnerships, and private equity investments. Public equity assets include
stocks and mutual funds, and safe assets include cash, bonds, retirement funds, and insurance. Net wealth and real estate exclude
personal residences.

40

including real estate, risky assets, public equity, and safe assets. Real estate assets include land,
residential and commercial buildings, and real estate partnerships. Risky assets include private
businesses, partnerships, and private equity investments. Public equity assets include stocks
and mutual funds, and safe assets include cash, bonds, retirement funds, and insurance.
The wealth for all congressmen increased from $6,180,330 to $7,605,058, from 2004 to
2012, with a small decline during the financial crisis. Large portions of their wealth are held in
real estate and private equity, with the majority of their wealth being held in public equity,
ranging from 33% to 56%, which is not unusual for wealthy individuals. Not surprisingly,
senators are much wealthier than House members, with the average senator reporting a mean
wealth of about $13,820,560 compared to $5,467,092 for House members.
3.3 Methodology
I examine the annual returns to congressmen’s wealth. To measure the performance of
their portfolios of total assets, I follow the standard approach in the empirical finance literature
and initially compare the wealth returns of congressmen to the risk adjusted market return.
Specifically I evaluate performance by using the Capital Asset Pricing Model (CAPM):
𝑅𝑝,𝑡 − 𝑅𝑡𝑓 = 𝛼 + 𝛽(𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝜀𝑝,𝑡
where 𝑅𝑝,𝑡 is the net wealth return for member 𝑖 in year 𝑡, 𝑅𝑡𝑚 is the return on the market
portfolio, and 𝑅𝑡𝑓 is the risk-free rate or return on US Treasury Bills. The key quantity of interest
in this panel regression is the intercept 𝛼, which measures the average abnormal return.

41

In addition, I evaluate performance of congressmen’s wealth portfolios by using the
Fama-French Three factor model:
𝑅𝑝,𝑡 − 𝑅𝑡𝑓 = 𝛼 + 𝛽𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝛽 𝑆𝑀𝐵 𝑆𝑀𝐵𝑡 + 𝛽𝐻𝑀𝐿 𝐻𝑀𝐿𝑡 + 𝜀𝑝,𝑡
where 𝑅𝑝,𝑡 is the net wealth return for member 𝑖 in year 𝑡, 𝑅𝑡𝑚 is the return on the market
portfolio, 𝑅𝑡𝑓 is the risk-free rate or return on US Treasury Bills, SMB is the difference between a
portfolio of small stocks and a portfolio of big stocks, and HML is the difference between a
portfolio of high book-to-market stocks and a portfolio of low book-to-market stocks. See Fama
and French (1993) for details on the construction of the factors. 19
Next, I test for asset selection ability and market timing abilities of the congressmen. A
number of studies have been proposed in the literature for measuring selectivity and market
timing ability.20 Here I consider two different regressions analyses. The first model was
proposed by Treynor and Mazuy ((1996), hereafter TM), which adds a quadratic term to the
CAPM equation (1). TM argued that if an individual can forecast market returns, he will adjust
his portfolios market exposure depending on the signal he receives about future market
returns. Thus, the portfolio will be described by the following equation:
2

𝑅𝑝,𝑡 − 𝑅𝑡𝑓 = 𝛼𝑝 + 𝛽𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝛾𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝜀𝑝𝑡
where the coefficients 𝛼𝑝 and 𝛾𝑝 represent selectivity and market timing skills, respectively. I

19

Data on the Fama-French three-factor model are obtained from Ken French’s website
http://mba.tuck.dartmouth.edu/pages/faculty/ken.french/
20
For examples see Treynor and Mazuy (1966), Jensen (1972), Fama (1972), Henriksson and Merton (1981),
Bhattacharya and Pfleiderer (1983), Chang and Lewellen (1984), among others.

42

also consider an extended version of this model, which includes the Fama-French size and value
factors to provide greater explanatory power than the basic model.
The second model I use to separate market timing and selectivity ability was proposed
by Henriksson and Merton ((1981), hereafter HM). The HM model is shown in the following
regression equation:
𝑅𝑝,𝑡 − 𝑅𝑡𝑓 = 𝛼𝑝 + 𝛽𝐷𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝛽(𝑈−𝐷)𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 )𝐷𝑈𝑡 + 𝜀𝑝𝑡
where 𝐷𝑈𝑡 is a dummy variable that is equal to 1 if the excess market return is positive and 0
otherwise. The coefficients 𝛼𝑝 and 𝛽(𝑈−𝐷)𝑝 represent selectivity and market timing skills,
respectively, where 𝛽(𝑈−𝐷)𝑝 equals the difference for portfolio 𝑝 between its up and down
market beta. As with the TM model, I also include an extended version of the HM model, which
includes the Fama-French size and value factors.
In addition to the basic CAPM, TM and HM models above, I test for selectivity and
market timing skills with an expanded version of each model which includes a measure of the
aggregate return to human capital. A significant criticism of the capital asset pricing model
(CAPM), which was first addressed by Mayers (1973) and Roll (1977), is that the traditional
measures of the market portfolio are incomplete since they do not consider returns to human
capital, which accounts for a large portion of wealth. Motivation for including measures of the
aggregate return to human capital to improve the definition of the market portfolio is given by
Fama and Schwert (1977), Jagannathan and Wang (1996) and Campbell (1996). More
specifically, Jagannathan and Wang (1996) find that the growth rate in aggregate labor income
is a priced risk factor and therefore, the human capital CAPM (HCAPM) significantly improves
43

upon the traditional CAPM. To incorporate human capital into CAPM, suppose the return on
the market portfolio of all assets is a weighted average of the return to financial wealth, 𝑅𝑡𝐹 ,
and the return to non-financial (human capital) wealth, 𝑅𝑡𝑁𝐹 . That is:
𝑅𝑡𝑚 = 𝜔𝑅𝑡𝑁𝐹 + (1 − 𝜔)𝑅𝑡𝐹
where 𝑅𝑡𝑚 is the return on the market portfolio and 𝜔 is the aggregate share of human wealth
in total wealth.
Some assumptions must be made about how to measure the return to human capital
since this is difficult to observe. Fama and Schwert (1977) and Jagannathan and Wang (1996)
assume that the expected return to human capital is constant. Under this assumption, the rate
of change in wealth is given by:

𝑅𝑡𝑁𝐹 =

𝐿𝑡 − 𝐿𝑡−1
𝐿𝑡−1

where 𝐿𝑡 is the difference between the National Income and Product Account (NIPA) of the USA
measure of annual total personal income and total dividends, normalized by total US
population, following Heaton and Lucas (2000)21. The return to labor is lagged one year to
account for the lags in the official reports of aggregate income. Using this approach, the return
to a particular congressman’s wealth, 𝑅𝑖𝑡 , can be expressed as:
𝑅𝑝𝑡 = 𝛼𝑝 + 𝛽𝐹 𝑅𝑡𝐹 + 𝛽𝑁𝐹 𝑅𝑡𝑁𝐹 + 𝜀𝑝𝑡
I proxy for the return to financial wealth with the value-weight return of all CRSP firms

21

From Table 2.2 in the National Income and Product Account of the USA published by the Bureau of Economic
Analysis, US Department of Commerce.

44

incorporated in the US and listed on the NYSE, AMEX, or NASDAQ.
In addition to the basic HCAPM model, I evaluate the returns to congressmen wealth by
including the three Fama-French factors in the HCAPM model:
𝑅𝑝𝑡 − 𝑅𝑡𝑓 = 𝛼𝑝 + 𝛽𝐹 𝑅𝑡𝐹 + 𝛽𝑁𝐹 𝑅𝑡𝑁𝐹 + 𝛽 𝑆𝑀𝐵 𝑅𝑡𝑆𝑀𝐵 + 𝛽𝐻𝑀𝐿 𝑅𝑡𝐻𝑀𝐿 + 𝜀𝑝𝑡
I also include the estimate for the return to human capital in the TM model:
2

𝑅𝑝,𝑡 − 𝑅𝑡𝑓 = 𝛼𝑝 + 𝛽𝑁𝐹 𝑅𝑡𝑁𝐹 + 𝛽𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝛾𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝜀𝑝𝑡
and the HM model:
𝑅𝑝,𝑡 − 𝑅𝑡𝑓 = 𝛼𝑝 +𝛽𝑁𝐹 𝑅𝑡𝑁𝐹 + 𝛽𝐷𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 ) + 𝛽(𝑈−𝐷)𝑝 (𝑅𝑡𝑚 − 𝑅𝑡𝑓 )𝐷𝑈𝑡 + 𝜀𝑝𝑡
The inclusion of human capital likely results in more accurate estimates than the standard TM
and HM models.
3.4 Empirical Results
The results from the CAPM and Fama-French three factor regressions are displayed in
Table 2. The table includes results for all congressmen, House members, and senators,
separately. To be included in my sample of wealth returns, congressman year returns must be
greater than -100%. The returns worse than -100% are dropped as outliers since many of these
largely negative numbers are due to congressmen reporting mortgages without also reporting
the value of the house and also due to large legal liabilities for congressmen fighting criminal
charges. For Congress as a whole, I find that members on average earn abnormal returns to
their wealth of 6.68% per year, using the CAPM model and 6.39%, using the Fama-French three45

Table 3.2: Fama-French Three Factor Model

All Congressmen

House

Senate

Alpha

Mktrf

6.6788***
(8.56)
6.3859***
(8.18)
6.9762***
(7.74)
6.6203***
(7.39)
5.5713***
(3.63)
5.5104***
(3.47)

0.2808***
(6.83)
0.2593***
(6.07)
0.2753***
(5.71)
0.2471***
(4.95)
0.2582***
(3.59)
0.2657***
(3.47)

SMB

0.2179*
(1.96)

HML

0.083
(0.98)

Adj. R2

N

0.0112

4249

0.0118
0.0102

0.2916**
(2.25)

0.0396
(0.41)

0.0111
0.0113

-0.0866
(-0.42)

0.2421
(1.39)

3376

839

0.0113

This table presents results for the CAPM and Fama-French Three-Factor models for all congressmen, House members, and
senators for the entire time period, 2004-2012. The dependent variable for both models is the net wealth return, 𝑅𝑝 , in excess of
the risk free rate, 𝑅𝑓 , observed at the end of each year. The intercept, 𝛼, measures the average annual abnormal return. The tstatistics, in parentheses, are adjusted for heteroscedasticity. ***, **, and * indicate significance at the 1%, 5%, and 10%, levels,
respectively

factor model. The results for the House and Senate are of similar strength. House members on
average earn positive abnormal returns to their wealth of 6.62%, when senators earn positive
abnormal returns of 5.5%. These

results are significant at the 1% level in all cases,

suggesting that congressmen are able to make superior returns on their wealth portfolios. The
t-statistics are adjusted for heteroscedasticity.
Table 3, panels A and B, display the means of the selectivity and timing values for all
congressmen, House members and senators, using the TM and HM models, respectively. As in
Table 2, the t-statistics are adjusted for heteroscedasticity. The TM model shows a significantly
positive mean alpha, the measure of selectivity, and an insignificant 𝛾𝑝 , the measure of market
timing ability. Alpha is strongly significant at the 1% level for all congressmen, House members,
and senators. The results of the HM model are similar. Alpha is significant at the 1% level

46

Table 3.3: Treynor-Mazuy and Henriksson-Merton Models
Panel A: Treynor-Mazuy Model
𝛼
All Congressmen

House

Senate

𝛽𝑝

𝛾𝑝

5.9287***
(5.35)
7.1737***
(5.23)
5.5384***
(4.35)
6.8875***
(4.29)
7.5596***
(3.30)
8.3865***
(3.42)

0.3042***
(6.22)
0.2230***
(3.21)
0.3201***
(5.56)
0.23486***
(2.87)
0.1970**
(2.39)
0.1330
(1.15)

0.0017
(0.89)
-0.002
(-0.68)
0.0032
(1.46)
-0.0007
(-0.19)
-0.0045
(-1.36)
-0.0072
(-1.54)

𝛼

𝛽𝐷𝑝

𝛽(𝑈−𝐷)𝑝

5.5257***
(4.51)
7.4615***
(4.33)
5.1173***
(3.65)
7.1700***
(3.56)
7.2698***
(2.83)
8.6856***
(2.77)

0.2234***
(3.47)
0.3084***
(3.75)
0.1833**
(2.45)
0.2722***
(2.83)
0.3453***
(2.88)
0.4124***
(2.86)

0.1451
(1.16)
-0.151
(-0.69)
0.2329
(1.59)
-0.077
(-0.30)
-0.2173
(-0.92)
-0.4497
(-1.17)

𝛽 𝑆𝑀𝐵

𝛽𝐻𝑀𝐿

Adj. R2
0.0112

0.3093*
(1.78)

0.0742
(0.86)

0.0117
0.0105

0.3224
(1.57)

0.0367
(0.37)

0.0108
0.0119

0.2544
(0.89)

0.205
(1.15)

0.0121

𝛽 𝑆𝑀𝐵

𝛽𝐻𝑀𝐿

Adj. R2

Panel B: HM model
All Congressmen

House

Senate

0.0113
0.3283*
(1.69)

0.0859
(1.01)

0.0117
0.0106

0.3477
(1.53)

0.0413
(0.42)

0.0108
0.011

0.2461
(0.73)

0.246
(1.41)

0.0115

This table reports results from the basic Treynor-Mazuy and Henriksson-Merton models, as well as the augmented models which
include the Fama-French factors SMB and HML. In the TM model the coefficients 𝛼𝑝 and 𝛾𝑝 represent the mean values of the
selectivity and market timing skills, respectively, for all congressmen, House members, and senators for the entire time period,
2004-2012. In the HM model the coefficient 𝛼𝑝 represent selectivity and 𝛽(𝑈−𝐷)𝑝 represents market timing skills. The t-statistics,
in parentheses, are adjusted for heteroscedasticity. ***, **, and * indicate significance at the 1%, 5%, and 10%, levels,
respectively

while, 𝛽(𝑈−𝐷)𝑝 , the measure for market timing, is insignificant in all cases. The results in Table 3
suggests that all congressmen, House members, and senators on average have superior ability
to select assets, but do not have market timing skills.

47

Table 4, Panel A presents results of the standard HCAPM model and the Fama-French
three factor model with the inclusion of human capital. Consistent with the results of the
traditional CAPM model, the results of the HCAPM model suggest that congressmen earn large
abnormal returns to wealth of 8.25%, while House members earn abnormal returns of 8.89%
and senators earn abnormal returns of 5.28%. The results in all cases are significant at the 1%
level. Also consistent with the results of the traditional Fama-French three factor model, the
Fama-French three factor model with the inclusion of human capital suggests that congressmen
earn large abnormal returns. Though the results are not as strong as the HCAPM model, they
are significant at the 1% level.
Table 4, Panels B presents results for the TM model with the inclusion of human capital.
Consistent with results of the standard TM model, the results of the TM model when the
average return to human capital is included suggest that all congressmen have asset selection
abilities. The TM model results also suggest both House members and senators have asset
selection abilities, but to not have market timing abilities. That is, alpha is significant at the 1%
level for all congressmen, House members, and senators. However, 𝛾𝑝 , is insignificant for all
congressmen, House members, and senators, suggesting they are not able to consistently time
the market.
Table 4, Panels C presents results for the HM model with the inclusion of human capital.
Consistent with all previous results, the results of the HM model when the return to human
capital is included suggest that all congressmen have asset selectivity abilities. Alpha is
significant at the 1% level for all congressmen and House members. However, the results of the

48

Table 3.4: HCAPM Model
Panel A: HCAPM Model
All Congressmen

House

Senate

𝛼

𝛽𝑝

𝛽 𝑁𝐹

8.2510***
(7.73)
7.9878***
(6.59)
8.8873***
(7.24)
8.2850***
(5.95)
5.2814**
(2.49)
6.2260***
(2.57)

0.2619***
(6.39)
0.2518***
(6.03)
0.2519***
(5.34)
0.2390***
(4.98)
0.2616***
(3.21)
0.2628***
(3.16)

0.8775**
(2.16)
0.8386*
(1.75)
1.0772**
(2.30)
0.8782
(1.59)
-0.1568
(-0.20)
0.3658
(0.39)

-0.1380
(-0.53)

0.2635
(1.44)

𝛼

𝛽𝑝

𝛾𝑝

𝛽 𝑁𝐹

𝛽 𝑆𝑀𝐵

8.4499***
(4.92)
9.3413***
(5.20)
8.1891***
(4.25)
9.0488***
(4.40)
8.7256***
(2.27)
9.7854***
(2.66)

0.2568***
(4.76)
0.1966***
(2.81)
0.2699***
(4.28)
0.2079**
(2.51)
0.1757*
(1.81)
0.1175
(1.02)

-0.0003
(-0.15 )
-0.0029
(-1.01)
0.0011
(0.44)
-0.0017
(-0.48)
-0.0054
(-1.32)
-0.0078
(-1.65)

0.9101**
(2.12)
0.9288**
(2.00)
0.9633**
(1.99)
0.9296*
(1.79)
0.4129
(0.43)
0.5943
(0.55)

𝛼

𝛽𝐷𝑝

𝛽(𝑈−𝐷)𝑝

𝛽 𝑁𝐹

8.0030***
(4.39)
9.7844***
(4.67)
7.8309***
(3.83)
9.5174***
(3.94)
7.9228*
(1.90)
10.0674**
(2.40)

0.2534***
(3.80)
0.3248***
(3.94)
0.2159***
(2.79)
0.2889***
(2.99)
0.3534***
(2.75)
0.4217***
(2.91)

0.0233
(0.16)
-0.2269
(-1.02 )
0.0994
(0.61)
-0.1551
(-0.60)
-0.2492
(-0.87)
-0.4914
(-1.27)

0.8426**
(1.99)
0.9332**
(2.01)
0.9288*
(1.95)
0.9442*
(1.82)
0.2184
(0.23)
0.5559
(0.52)

𝛽 𝑆𝑀𝐵

Adj. R2

𝛽 𝐻𝑀𝐿

0.0121
0.1035
(0.78)

0.1291
(1.38)

0.0123
0.0114

0.1731
(1.14)

0.0870
(0.80)

0.0115
0.0101
0.0103

Panel B: Treynor-Mazuy Model
All Congressmen

House

Senate

𝛽 𝐻𝑀𝐿

Adj. R2
0.0118

0.2283
(1.26)

0.1208
(1.34)

0.0123
0.0112

0.2430
(1.15)

0.0826
(0.80)

0.0113
0.011

0.1988
(0.62)

0.2367
(1.24)

0.0114

𝛽 𝑆𝑀𝐵

𝛽 𝐻𝑀𝐿

Adj. R2

Panel C: HM model
All Congressmen

House

Senate

0.0118
0.2565
(1.28)

0.1387
(1.53)

0.0123
0.0112

0.2772
(1.19)

0.0940
(0.91 )

0.0113
0.0099

0.1987
(0.54)

0.2789
(1.49)

0.0107

This table reports results from the CAPM, TM and HM models with the inclusion of the human capital factor, as well as the
augmented models which include the Fama-French factors SMB and HML. In the TM model the coefficients 𝛼 and 𝛾𝑝 represent
the mean values of the selectivity and market timing skills, respectively. In the HM model the coefficient 𝛼𝑝 represent selectivity
and 𝛽(𝑈−𝐷)𝑝 represents market timing skills. The models are used for all congressmen, House members, and senators separately
for the entire time period, 2004-2012. The t-statistics, in parentheses, are adjusted for heteroscedasticity. ***, **, and * indicate
significance at the 1%, 5%, and 10%, levels, respectively

49

human capital HM model are significant at the 10% level for senators and significant at the 5%
level when the Fama-French factors are included. 𝛽(𝑈−𝐷)𝑝 , the measure of market timing
ability is insignificant in all cases.
The results in Tables 3 and 4 suggest that all congressmen, as well as both Houses of
Congress individually, on average have superior asset selectivity abilities but do not have
market timing skills. That congressmen are able to increase their wealth through their ability to
select profitable assets is consistent with previous research. Eggers and Hainmueller’s (2014)
find that members invest disproportionately in local firms and campaign contributors and that
these investments outperform their non-local and non-connected investments, as well as the
market as a whole.22 These results suggest that politicians may use information gained through
their interactions with local firms and campaign contributors when determining their stock
portfolios or are able to influence legislation that impacts corporations in their portfolios
through, for example, earmarks or granting government contracts. Tahoun (2014) finds
evidence that firms with a strong ownership-contribution relationship, measured by
congressmen’s stock ownership and firms’ campaign contributions, receive more government
contracts. In addition, Tahoun and van Lent (2010) find that US congressmen’s equity
ownership is positively correlated with their voting in favor of legislative proposals to bail out
the financial sector. Taken together, these studies support the hypotheses that congressmen
use the information they gain through their positions of power to carefully select their
investment and/or influence legislation to benefit their portfolios.
22

This local bias is approximately twice as large as what has been shown for other individuals (Ivkovic and
Weisbenner, 2005) and approximately ten times larger than that found for professional money managers (Coval
and Moskowitz, 1999)

50

3.5 Powerful Congressmen
If abnormal returns to wealth are a result of members of Congress granting advantages
to companies or communities in which they own assets, then I would expect returns to wealth
to depend on the political power of congressmen. The political science literature suggests that
congressional committees play a key role in establishing strong politician-firm relationships due
to repeated interactions (Kroszner and Stratmann, 1998). In addition, the literature shows that
committees do not have equal ability to award benefits to their connections (Edwards and
Stewart, 2006) and a politician’s power affects the distributions of government benefits
(Roberts, 1990). Thus, it is likely that powerful committee members earn larger returns to
wealth than non-committee members, due to greater access to private information and greater
abilities to bestow benefits to their connections.
To rule out the possibility that powerful politicians are driving the results in the previous
section, I separately measure the returns to wealth for powerful politicians. Table 5 replicates
the CAPM, TM and HM models, with and without the inclusion of human capital, for members
of Congress who serve on powerful committees. Following Eggers and Hainmueller (2014),
powerful committees in the House are defined as Rules, Appropriations, Ways and Means, and
Commerce and in the Senate they are defined as Appropriations, Finance, and Commerce. The
results in Table 5, Panel A are not supportive of the idea that positive abnormal returns to
wealth are being driven by the most powerful members. That is, the Fama-French three-factor
model shows that powerful congressmen earn abnormal returns to wealth of 5.65%, significant
at the 1% level, while the results for “All Congressmen” in Table 3 suggest congressmen on

51

Table 3.5: Returns to Wealth for Powerful Members of Congress
Panel A: Fama-French three-factor model with and without human capital
All
Congressmen

House

Senate

𝛼

𝛽𝑝

𝛽 𝑆𝑀𝐵

𝛽 𝐻𝑀𝐿

5.6526***
(5.17)
7.8115***
(4.25)
5.8730***
(4.55)
7.0765***
(3.49)
5.0519**
(2.41)
9.2115**
(2.34)

0.2196***
(3.50)
0.2122***
(3.39)
0.2077***
(2.69)
0.2039***
(2.64)
0.2341**
(2.20)
0.2172**
(2.07)

0.2676
(1.60)
0.1084
(0.53)
0.4151*
(1.96)
0.3252
(1.32)
-0.1520
(-0.58)
-0.4488
(-1.24)

0.2924**
(2.52)
0.3548***
(2.85)
0.1386*
(1.84)
0.2880*
(1.96)
0.3817
(1.77)
0.5151**
(2.15)

𝛽 𝑁𝐹

1.0811*
(1.65)

Adj. R2

N

0.0136

1746

0.0144
0.0141

0.5932
(0.81)

1227

0.0137
0.0086

2.1564
(1.56)

505

0.0128

Panel B: TM model with and without human capital
All
Congressmen

House

Senate

𝛼

𝛽𝑝

𝛾𝑝

𝛽 𝑆𝑀𝐵

𝛽 𝐻𝑀𝐿

4.3630**
(2.15)
6.7866**
(2.58)
4.0942
(1.58)
5.2615*
(1.68)
5.2948*
(1.70)
10.4143**
(2.14)

0.2783***
(2.73)
0.2537**
(2.49)
0.2889**
(2.22)
0.2773**
(2.13)
0.2231
(1.48)
0.1684
(1.14)

0.0032
(0.74)
0.0023
(0.51)
0.0044
(0.78)
0.0040
(0.69)
-0.0006
(-0.10)
-0.0027
(-0.44)

0.1097
(0.38)
0.0070
(0.02)
0.1965
(0.51)
0.1461
(0.37)
-0.1225
(-0.34)
-0.3291
(-0.80)

0.3113***
(2.60)
0.3645***
(2.87)
0.2812**
(1.96)
0.3056**
(2.03)
0.3783*
(1.70)
0.5045
(2.07)

𝛽 𝑁𝐹

1.0195
(1.55)

Adj. R2

N

0.0133

1746

0.0140
0.0140

0.4835
(0.66)

1227

0.0134
0.0066

2.2279
(1.60)

0.0111

𝛽 𝑁𝐹
0.5247
(0.89)
1.0377
(1.59)
0.2146
(0.32)
0.5322
(0.74)
1.1812
(1.00)
2.1717
(1.57)

Adj. R2
0.0110

505

Panel C: HM model with and without human capital
All
Congressmen

House

Senate

𝛼
5.1342**
(2.02)
6.8465**
(2.27)
3.4876
(1.26)
5.7153
(1.58)
9.1417
(1.64)
9.5493*
(1.74)

𝛽𝐷𝑝

𝛽(𝑈−𝐷)𝑝

0.1381
(1.43)
0.1720
(1.38)
0.0759
(0.66)
0.1476
(0.94)
0.2906
(1.63)
0.2315
(1.19)

0.2428
(1.25)
0.1238
(0.37)
0.4204*
(1.86)
0.1741
(0.41)
-0.2371
(-0.62)
-0.0437
(-0.09)

𝛽 𝑆𝑀𝐵

𝛽 𝐻𝑀𝐿

0.0203
(0.06)

0.3526***
(2.83)

0.2010
(0.47)

-0.4178
(-0.88)

0.2854*
(1.94)

0.5162**
(2.17)

N
1746

0.0139
0.0116

1227

0.0131
0.0048

505

0.0108

This table reports results from the Fama-French three-factor, TM and HM models with and without the inclusion of the human
capital factor. In the TM model the coefficients 𝛼 and 𝛾𝑝 represent the mean values of the selectivity and market timing skills,
respectively. In the HM model the coefficient 𝛼𝑝 represent selectivity and 𝛽(𝑈−𝐷)𝑝 represents market timing skills. 𝛽𝑁𝐹
measures the return to human capital. These models are used for all congressmen, House members, and senators separately for
the entire time period, 2004-2012. ***, **, and * indicate significance at the 1%, 5%, and 10%, levels, respectively.

52

average earn abnormal returns to wealth of 6.39%. The results are similar for House members
and senators. In both cases, the alpha estimates for all House members and all senators are
slightly greater than the results for powerful House members and senators. The interpretation
of the results in Table 5 is similar when the human capital factor is included. Taken together,
these results suggest, that if anything, the congressmen who do not serve on powerful
committees earn greater returns to wealth than those who do serve on powerful committees.
In Table 5, Panels B and C I examine whether powerful Congress members’ returns to
wealth are due to selectivity or market timing abilities. Consistent with the results in Tables 3
and 4, the results in Table 5 suggest powerful congressmen have superior abilities to select
profitable assets. However, these results are weaker than those for the entire sample of
congressmen, House members, and senators, providing further evidence against the
hypotheses that powerful congressmen may be driving the results in Tables 3 and 4.
3.6 Conclusion
A substantial body of research in political economy supports the idea that politicians
may get rich by investing on their private government information by examining public
connections between politicians and firms (Tahoun, 2014; Correia, 2009; Yu and Yu, 2006;
Tahoun and van Lent, 2010; Jayachandran, 2006; Goldman, Rocholl, and So, 2009; Eggers and
Hainmueller, 2014). By focusing on connections that are subject to public scrutiny, results are
subject to a downward bias due to congressmen’s concerns about their reputations. In this
chapter, I investigated whether congressmen act corruptly by focusing on returns to their total
wealth rather than only focusing on their public equity investments. The use of total wealth

53

likely provides a better estimate of the extent to which congressmen use private government
information for personal gain, since total wealth measures capture gains to real estate/ land
ownership, private equity investments, and small business ownership. These assets have a
more private nature than public equity, thus providing greater opportunities for congressmen
to behave unethically.
I find evidence that congressmen’s use of private information for personal gain may be
more widespread than previous research suggests. Using the Fama-French Three-Factor model
and Jagannathan and Wang’s (1996) human capital models, I find that congressmen earn large
abnormal returns to wealth of 6.39% per year on average, thus providing further support to the
public perception that congressmen may unethically gain from their positions of power and
proving, once again, that incentives are important drivers of behavior. While Eggers and
Hainmueller’s (2013) results that congressmen’s stock portfolios have mediocre performance
suggest that public disclosure of their financial assets has discouraged the use of private
information for gains in public markets, the results in this chapter suggest that congressmen
may have more private methods of using their private information for personal gain.
In addition to testing for abnormal returns to wealth, I also examine whether
congressmen earn large returns as a result of their ability to select profitable assets or
successfully time the market. Using two models, the first proposed by Treynor and Mazuy
((1996) and the second proposed Henriksson and Merton ((1981), I find that the large returns
to wealth are due to superior asset selectivity rather than market timing skills. These results
are consistent with previous literature which suggests congressmen carefully select where to

54

invest based on their government positions and may attempt to influence legislation depending
on their portfolio holdings (Eggers and Hainmueller, 2014; Tahoun, 2014; Tahoun and van Lent;
2010).
The results of this chapter warrant further research. There is evidence that
congressmen’s local public equity investments perform very well (Eggers and Hainmueller,
2014). This is likely the case for congressmen’s private investments as well. While this study
indirectly tests for whether congressmen earn large returns to their wealth due to ownership of
assets in opaque markets, it may be possible to test this more directly by examining whether
there is an association between the earmarks congressmen insert in legislation and increases in
their wealth.

55

BIBLIOGRAPHY

56

BIBLIOGRAPHY

Barber, B. M. and T. Odean. 2000. “Trading Is Hazardous to Your Wealth: The Common Stock
Investment Performance of Individual Investors.” The Journal of Finance 55(2): 773–806.
Barber, B. M., Y. T. Lee, Y. J. Liu, and T. Odean. 2009. “Just How Much Do Individual
Investors Lose by Trading?” Review of Financial Studies 22(2): 609–632.
Barberis, N. and R. Thaler. 2003. “A Survey of Behavioral Finance.” Handbook of the
Economics of Finance 1: 1053–1128.
Barro, Robert. 1973. “The Control of Politicians: An Economic Model.” Public Choice 14 (1): 1942
Besley, Timothy, and Anne Case. 1995. “Does Electoral Accountability Affect Economic Policy
Choices? Evidence from Gubernatorial Term Limits.” Quarterly Journal of Economics, 110(3):
769–98.
Bessen, James E., 2016. "Accounting for Rising Corporate Profits: Intangibles or Regulatory
Rents?." Boston Univ. School of Law, Law and Economics Research Paper 16-18.
Bhattacharya S. and P. Pfleiderer. 1983. “A Note on Performance Evaluation”, Technical Report
714 Stanford, Calif: Stanford University, Graduate School of Business.
Campbell, John Y.. 1996. “Understanding Risk and Return,” Journal of Political Economy, 104,
298-345.
Chang E. and W. G. Lewellen. 1984. “Market Timing and Mutual Fund Investment
Performance”, Journal of Busi-ness, Vol. 57, No. 1, pp. 57-72.
Claessens, S., Feijen, E., Laeven, L., 2008. “Political connections and preferential access to
finance: The role of campaign contributions.” Journal of Financial Economics 88, 554–580.
Cohen, L., Coval, J.D. and Malloy, C., 2010. Do powerful politicians cause corporate downsizing?
(No. w15839). National Bureau of Economic Research.
Correia, M., 2009. Political connections, SEC enforcement and accounting quality. Unpublished
working paper. London Business School.
Daniel, K., M. Grinblatt, S. Titman, and R. Wermers. 1997. Measuring Mutual Fund Performance
with Characteristic-based Benchmarks. Journal of Finance 52:1035–58.

57

Edwards, Keith, and Charles Stewart III., 2006. "The value of committee assignments in
Congress since 1994." Annual Meeting of the Southern Political Science Association, Atlanta, GA
Eggers, A. and J. Hainmueller. 2013. “Capitol Losses: The Mediocre Performance of
Congressional Stock Portfolios.” The Journal of Politics 75(2): 535–551
Eggers, Andrew C., and Jens Hainmueller., 2014. "Political capital: Corporate connections and
stock investments in the US congress, 2004-2008." Quarterly Journal of Political Science: 201226.
Faccio, M. 2006. “Politically Connected Firms.” American Economic Review 96(1): 369–386.
Faccio, M., Masulis, R., McConnell, J., 2006. Political connections and corporate bailouts. The
Journal of Finance 61, 2597–2635.
Fama E., 1972. “Components of Investment Performance”, Journal of Finance, Vol. 27, No. 2,
pp. 551-567.
Fama, E. F., and K. R. French., 1993. "Common Risk Factors in Returns on Stocks and Bonds."
Journal of Financial Economics, 33, 3-56.
Fama, E. F., and W. G. Schwert, 1977, “Human capital and capital market equilibrium,” Journal
of Finance, 51, p. 55-84.
Ferraz, Claudio, and Frederico Finan. 2009. “Electoral Accountability and Corruption: Evidence
from the Audits of Local Governments.” American Economic Review, forthcoming
Ferejohn, John. 1986. “Incumbent Performance and Electoral Control.” Public Choice 50 (1-3):
5-26
Fisman, R. 2001. “Estimating the Value of Political Connections.” American Economic Review
91(4): 1095–1102.
Henriksson R. and R. Merton, 1981. “On Market Timing and Investment Performance”, Journal
of Business, Vol. 57, No. 4, pp. 513-534.
Hoechle, D., M. Schmid, and H. Zimmermann. 2009. “A Generalization of the Calendar Time
Portfolio Approach and the Performance of Private Investors.” WWZ Working Paper.
Ivkovi´c, Z. and S. Weisbenner. 2005. “Local Does as Local Is: Information Content of the
Geography of Individual Investors’ Common Stock Investments.” The Journal of Finance 60(1):
267–306.

58

Jagannathan, R. and Zhenyu Wang, 1996, “The Conditional CAPM and the Cross Section of
Expected Returns,” Journal of Finance, 51, 3-54.
Jeng, L. A., Metrick, A., & Zeckhauser, R. 2003. “Estimating the returns to insider trading: A
performance-evaluation perspective.” Review of Economics and Statistics, 85(2), 453-471.
Jensen M., 1972. “Optimal Utilization of Market Forecasts and Evaluation of Investment
Performance”, G.P. Szego and K. Shell eds, Mathematical Methods of Investment and Finance,
North – Holland Amsterdam.
Johnson, S. and T. Mitton. 2003. “Cronyism and Capital Controls: Evidence from Malaysia.”
Journal of Financial Economics 67(2): 351–382.
Khwaja, A. I. and A. Mian. 2005. “Do Lenders Favor Politically Connected Firms? Rent Provision
in an Emerging Financial Market.” The Quarterly Journal of Economics 120(4): 1371–1411.
Lenz, Gabriel, and Kevin Lim, 2009, “Getting Rich(er) in Office? Corruption and Wealth
Accumulation in Congress”, Working Paper.
Mayers D., 1973, “Nonmarketable Assets and the Determination of Capital Asset Prices in the
Absence of Riskless Rate,” Journal of Business, 46, p. 258-267.
Odean, T. 1999. “Do Investors Trade Too Much?” American Economic Review 89(5):1279–1298.
Querubin, Pablo and James Snyder, Returns to U.S. Congressional Seats in the Mid-19th
Century, in The Political Economy of Democracy, Enriqueta Aragones, Carmen Bevia, Humberto
Llavador, and Norman Schofield (eds.), 2009.
Roberts, B.E., 1990. A dead senator tells no lies: Seniority and the distribution of federal
benefits. american Journal of Political science, pp.31-58.
Roll, R., 1977, “A Critique of the Asset Pricing Theory’s Tests,” Journal of Financial Economics, 4,
p. 129-176.
Schweizer, P. 2011. Throw Them All Out: How Politicians and Their Friends Get Rich Off Insider
Stock Tips, Land Deals, and Cronyism That Would Send the Rest of Us to Prison. Houghton
Mifflin Harcourt (HMH).
Tahoun, A., 2014. The role of stock ownership by US members of Congress on the market for
political favors. Journal of Financial Economics 111 (1), 86-110.
Treynor J.L. and J. Mazuy, 1966. “Can Mutual Funds Outguess the Market?” Harvard Business
Review, Vol. 44, No. 4, pp. 131 – 136.
59

Yu, F., & Yu, X. 2012. Corporate lobbying and fraud detection. Journal of Financial and
Quantitative Analysis, 46(06), 1865-1891.
Ziobrowski, A. J., P. Cheng, J. W. Boyd, and B. J. Ziobrowski. 2004. “Abnormal Returns from the
Common Stock Investments of the US Senate.” Journal of Financial and Quantitative Analysis
39(4): 661–676.
Ziobrowski, A. J., J. W. Boyd, P. Cheng, and B. J. Ziobrowski. 2011. “Abnormal Returns From the
Common Stock Investments of Members of the US House of Representatives.” Business and
Politics 13(1): 4.

60