This document summarizes a research paper analyzing campaign spending in Brazil. It discusses how previous research by David Samuels found that, unlike in the US, campaign spending is equally important for both incumbents and challengers in Brazil due to the weaker incumbency advantage. The document aims to expand on Samuels' work by analyzing how candidate quality impacts the importance of campaign spending. It will test whether campaign spending matters more for low-quality challengers compared to high-quality incumbents. The document outlines the methodology, which includes performing statistical tests on campaign spending data from Brazil to determine spending thresholds for winning a seat. It also reviews previous literature on comparing campaign finance in Brazil and the US.
The Presidential Race: A look at election speeches and what personality wins
Paper
1.
The Importance of Campaign Spending For
Incumbents and Challengers in the Brazilian
Chamber of Deputies:
The Candidate Quality Argument
By Joel Ames
POLS 496
Latin American Politics
Professor Juan Pablo Micozzi
2. 2
Abstract
Campaign spending in elections has been a popular subject among scholars of
political science in the United States. Not much about campaign spending has been
studied in Latin America though. One scholar has done an excellent job of analyzing
campaign spending in Brazil. Professor David Samuels argues that both incumbents and
challengers in the Brazilian Chamber of Deputies value campaign spending equally
because of the weak incumbency advantage. In this piece, I analyze David Samuels’
campaign spending data from the 1989 Brazilian Chamber of Deputies election. I
hypothesize that though Samuels’ theory is logical, when isolating the candidate quality
variable in the data, the importance of campaign spending to certain candidates will be
impacted. In other words, when high quality incumbents compete against low quality
challengers, the importance of campaign spending will be greater for challengers.
3. 3
Campaign spending in elections has been a popular subject for scholars of
political science in recent years. How much does campaign spending really matter for
election outcomes? Many scholars have attempted to answer this question in the United
States. However, there are very few pieces pertaining to campaign spending in Latin
American elections. One scholas in particular goes into great detail about campaign
spending in Brazilian elections. David Samuels has made great strides in explaining the
importance of campaign spending in Brazil. This paper will expand on and analyze his
previous work dealing with differences between incumbent and challenger campaign
spending in Brazil and the United States. It will also use David Samuels’ model to delve
further into the importance of candidate quality in both United States and Brazilian
elections, to see if there is a difference between high quality candidates and low quality
candidates while relating it to their individual campaign spending.
The piece that this paper will expand on and analyze further is David Samuels’
“Incumbents and Challengers on a Level Playing Field: Assessing the Impact of
Campaign Finance in Brazil”. In this piece, David Samuels hypothesizes that unlike the
United States, campaign spending matters equally for both challengers and incumbents in
Brazil. While the United States has an extreme incumbency advantage in the House of
Representatives, Brazil’s incumbency advantage in the Chamber of Deputies is much
weaker (Samuels 2001). Therefore, candidates who are challengers and incumbents in
Brazil value campaign spending equally for the purpose of gaining name recognition and
political clout.
Using David Samuels’ data from the piece, I want to pinpoint the percentage of
campaign spending in each Brazilian district that a candidate must spend to gain a
4. 4
positive chance of gaining a seat. In other words, the paper’s first goal is to pinpoint a
threshold for a large district in Brazil. This threshold will show the percentage of
campaign funding a candidate must achieve to have a better than 50-50 shot of winning a
seat. This will add to the results that David Samuels found in his paper. This paper will
not only perform the test on all the candidates from the data sample, but it will also focus
in on the candidates with high and low quality grades that were assigned in Samuels’
piece. So the threshold test will be performed on candidates that received the highest and
lowest scores of quality in a large district.
Though scholars examining the United States House of Representatives have
already determined this threshold, the paper will use David Samuels’ empirical model to
set a basis for a future theory for incumbents and challengers in the United States. By
narrowing the sample to those candidates considered “high quality”, this paper will
attempt to determine whether the incumbency advantage theory David Samuels presents
changes when looking at only high quality candidates from Brazil and the United States.
The results may show that the apparent difference in incumbency advantage between
Brazil and the United States may not apply for high quality candidates that are
challengers. In other words, campaign spending may be more important for incumbents
in the United States when facing a high quality candidate. In the same sense, campaign
spending may not be as equally important when a high quality candidate is involved in
Brazil.
The paper consists of four sections. The first section of this paper starts by further
examining David Samuels’ piece on campaign spending for incumbents and challengers
in Brazil. His section of Candidate quality will be examined further as well. Since, this
5. 5
paper attempts to draw a comparison between high quality candidates and campaign
spending in Brazil and in the United States, literature will be reviewed on why the
countries are good to compare. Previous literature has determined that with Brazil’s
open-list proportional representation method of elections, candidates do indeed foster a
personal vote. This makes it easier to compare to the United States, which in the
electoral system also fosters a personal vote. Previous literature will be examined to also
review the work on campaign spending in the United States. Peter Jacobson’s piece
identifying the amount a challenger must spend to compete for a seat in the House of
Representatives will be reviewed.
After reviewing all the literature on campaign spending in Brazil and the United
States, a hypothesis is then formed in the next section. It is expected that when
examining the factor of candidate quality further, that the strength of the incumbency
advantage in both Brazil and the United States will change. David Samuels makes an
extremely valid argument when determining that campaign spending matters equally for
both incumbents and challengers in Brazil, while the incumbency advantage in the United
States requires that incumbents worry about spending less than challengers. This paper’s
purpose is not to challenge his findings. Rather it is to expand on his variable of
candidate quality, while finding the threshold to achieve a positive chance of winning a
seat in Brazil based on the quality of the candidate.
The next section of the paper includes the testing and analyzing of the data. A
logistic regression test is performed on David Samuels’ data from his piece “Incumbents
and Challengers on a Level Playing Field: Assessing the Impact of Campaign Finance in
Brazil”. Then the same tests will be performed again, however, this time only candidates
6. 6
with the highest and lowest candidate quality rating will be examined. Next a “Pr-value”
test is performed on the district with the largest magnitude in the data (Sao Paolo) to
pinpoint the percentage of campaign spending a candidate must have to achieve a
positive chance of winning a seat. Finally, summary statistics will be drawn from the Sao
Paolo district, and they will be tied in to the “Pr-value” test. These tests should yield
results that reveal a not so equal playing field when it comes to campaign spending in the
Brazilian Chamber of Deputies. The results will be compared to see if the incumbency
advantage theory David Samuels presents still holds for Brazil and a new theory will be
presented for further research into the United States. Finally, the results are displayed
and interpreted, and a conclusion is drawn based on the findings.
Literature Review
David Samuels’ piece on incumbent and challenger campaign spending in Brazil
has been used as a basis for all research questions on campaign spending in Latin
America. It has also done something that many scholars have not done. The piece looks
to compare campaign spending in Brazil to campaign spending in the United States: “Yet
despite the potentially critical role of campaign finance, research on this issue remains
largely confined to the US. Surprisingly few studies explore campaign finance in
comparative perspective” (Samuels 2000). Samuels has written many pieces exploring
the similarities and differences between campaign spending in Brazil and the United
States. The one this paper will evaluate the most is the one dealing with the idea of a
level playing field with campaign spending for incumbents and challengers for Brazil.
7. 7
He argues that “Brazilian incumbents and challengers translate money into votes at equal
rates. This contrasts with a prominent claim about U.S. House elections---that
incumbents translate money into votes less effectively than challengers” (Samuels 2001).
Before jumping into his theory and findings more, it is important to establish why Brazil
and the United States can be compared without any outside variables that would taint the
research. Most electoral systems in Latin America are significantly different from the
United States; however, Brazil’s electoral system has a necessary component similar to
the United States’.
Scholars of political science have determined that the Brazilian Chamber of
Deputies is one of the most studied chambers of government, behind the U. S. House of
Representatives (Jones 2002). Obviously this paper is analyzing work by David Samuels,
who is one of the main contributors to these forms of research. Mark Jones explains why
scholars are studying the Brazilian Chamber of Deputies and the U.S. House of
Representatives more than any other legislative institution:
“One prominent explanation for the relatively large-scale study of the Brazilian
Chamber of Deputies is the candidate-centered nature of legislators, which makes
the application of theories developed for the study of U.S. Congress more easily
transferable than to party-centered countries and also makes the individual
legislator the most relevant unit of analysis, as in the U.S. Congress” (Jones
2002).
In other words, both the United States House of Representatives and the Brazilian
Chamber of Deputies foster the personal vote. Candidates will individually go out to
gain name recognition and clout to win a seat in office. Barry Ames, in The Deadlock of
8. 8
Democracy in Brazil, argues that unlike many Latin American countries, the Brazilian
Chamber of Deputies is an outlier when it comes to candidate orientation and behavior.
Candidates of the Brazilian Chamber of Deputies are much more individualistic and free
from party constraints (Ames 2001). This is a direct result from the combination of
Brazil’s open-list proportional representation system, the large districts magnitudes
present in the electoral system, and their nomination rules. Anyone can run in Brazil’s
electoral system. Like the United States, these nominees raise and spend their own
campaign funds and there is no set restriction on how much they can spend (Samuels
2001). Therefore it fosters this idea of a personal vote. Unknown candidates must gain
name recognition through individualistic campaigning which strays away from party
affiliation.
It has been established that both the Brazilian Chamber of Deputies and the
United States House of Representatives hold elections that foster the personal vote. This
is why they can be compared without any outside variable that will affect the validity of
any research or tests. Looking into Samuels’ piece on the level playing field for
challengers and incumbents in Brazil when it comes to campaign spending, Samuels first
points out how important money is to elections. He then goes on to explore the idea of a
campaign-spending limit in Brazil. Most scholars argue that that in the United States,
spending limits would create a protection agent for incumbents, due to the fact that
incumbents experience diminishing returns after spending too much anyway. Samuels
finds evidence that this isn’t true in Brazil, and describes four factors that contribute to
Brazil’s weak incumbency advantage: “holding a seat in the Chamber provides little
political payoff in terms of name recognition, the best incumbents often choose not to run
9. 9
again, challengers are often more prominent than incumbents, and the electoral system
undermines incumbents’ self-promotional efforts” (Samuels 2001). Samuels’ statistical
evidence that campaign spending matters equally for both incumbents and challengers in
Brazil is extremely valid. The one thing that this paper is trying to tweak is analyzing
further his variable of candidate quality. Similar notions have been presented towards
research on campaign spending in the United States House of Representatives.
Gary Jacobson, from the Department of Political Science at UC San Diego has
provided data regarding campaign spending in the United States House of
Representatives. In fact, his findings detailing the threshold needed for a candidate to
gain a positive chance of winning a seat in the House have inspired the foundation of this
paper. However, as stated earlier, this particular research is being designed to thoroughly
control for the candidate quality variable.
“Others argue that Jacobson does not do this in his research. Unfortunately
Jacobson's data analysis either neglects the direct effect of candidate quality
entirely or severely underestimates the influence of quality due to poor
measurement. Integrating challenger quality, its interaction with challenger
spending, and the effects of incumbent spending into Jacobson’s framework
improves the predictive accuracy of the model and contributes to our
understanding of the House vote” (Green 1988).
So while campaign spending may matter more for challengers and less for incumbents in
the large spectrum of the electoral system of the House of representatives, Donald Green
and Jonathan Krasno argue that this is not always the case when a quality rated
challenger is involved. Once a high quality challenger is involved, the importance of
10. 10
campaign spending may become more equal for challengers and incumbents. This is
exactly what this paper is trying to assess in Brazil. By focusing in on campaign
spending by only high quality challengers and incumbents, there may be an ambiguity
towards David Samuels’ theory that a level playing field exists in Brazil.
Hypothesis and Theory
As established earlier, money matters in getting elected. What is undoubtedly a
heated debate among scholars is left somewhat unanswered. Who does campaign
spending matter more too? Many scholars have done significant research on this
particular topic, including David Samuels, who focuses in on the Brazilian Chamber of
deputies, and Gary Jacobson, who analyzes the United States House of Representatives.
Both put forward excellent research on each legislative entity. However, this paper’s
goal is to analyze further into their data and focus in on high quality challengers for both
the House of Representatives and the Brazilian Chamber of Deputies. Again, this is not a
challenge against their theories. Rather, it is a more in depth look into the candidate
quality variable and the effect on campaign spending.
David Samuels argues that campaign spending is equally important for both
challengers and incumbents in the Brazilian Chamber of Deputies. However, I
hypothesize that when there is a significant difference in candidate quality between who
is running in each district, money will matter more to certain candidates. For instance,
when a challenger that is given a quality rating of one, according to David Samuels’ data,
11. 11
is competing with an incumbent with a candidate quality rating of three (the highest
rating), money will matter more to the challenger. In a similar sense, when a challenger
has a quality rating of three and is competing against an incumbent with a rating of one,
money will matter more to the incumbent. To summarize, candidate quality does have an
impact on how important campaign spending is to incumbents and challengers.
Therefore, the importance of campaign spending in situations where there are candidates
with very different campaign qualities will not be as equal as past research has shown.
This can also be predicted towards the United States House of Representatives.
In Gary Jacobson’s piece on campaign spending in the United States House of
Representatives, he argues that campaign spending matters significantly less for
incumbents, who will see diminishing returns after spending over a certain amount.
Campaign spending for challengers, on the other hand, matters a great deal to gain name
recognition and momentum. Similar to the hypothesis made towards campaign spending
in Brazil, I hypothesize that when isolating situations where both candidate’s quality
rating is very different, the importance of campaign spending will be impacted. In other
words, when there is a high rating in quality for a challenger competing against an
incumbent with a low quality rating, money will matter equally, if not more for the
incumbent. Though Gary Jacobson’s data will not be fully tested in this piece, the
findings will help set up for a comparison in research that is later to come. Essentially,
candidate quality will have an impact on the importance of campaign spending for
incumbents and challengers in the Brazilian Chamber of Deputies and the United States
House of Representatives. I can’t stress enough that these claims are not challenging the
theories of David Samuels and Gary Jacobson. Rather, they are attempting to isolate the
12. 12
candidate quality variable to see its true impact on the importance of campaign spending
for candidates.
Data and Methods
The first test that is performed in this research deals with a logistic regression on
David Samuels’ previous data, while adding a “yes_no” variable. A zero was assigned to
the candidates within the top portion of their district based on vote percentage. Since the
district names were not distinguished in the data, I assumed that, for instance, in districts
with a magnitude of eight, the top eight candidate’s vote percentages were assigned a
zero, meaning they gained a seat. However, there were multiple districts with
magnitudes of eight (11 total). Therefore, the top eighty-eight candidate’s vote
percentages received a zero. After running this test, the same test is repeated, only this
time the data is isolated. I use filters to isolate the candidate quality variable. This
occurred in two ways. First, I tested all of the data from the eleven districts that had a
magnitude of eight. A filter was put on to exclude all of the candidates that received a
candidate quality rating of two. The reason this filter was put on was to create a test of
extreme cases: the highest quality rating and the lowest quality rating. In other words,
this isolated the extreme cases where either an incumbent or a challenger had a great
advantage in candidate quality in the 1989 elections. Only candidates with a quality
rating of 1 and 3 are tested. In doing so, I hope to pinpoint the extreme cases where
candidate quality definitely matters.
13. 13
The next test in the research attempts to pinpoint percentages that explain how
much money a candidate needs to gain a seat in the largest district of the Brazilian
Chamber of Deputies; Sao Paolo. In doing so, the test gives out percentages based on the
amount of campaign funding received by the candidates. The first percentage describes
the chance the candidate has of gaining a seat based on their funding received. The
second percentage is just the first percentage subtracted from one hundred, which
describes the chance a candidate won’t gain a seat. David Samuels’ data is used,
however, a few tweaks are made to run the test Using the new yes_no variable, a logistic
regression test is used on David Samuel’s tweaked data. Then the Pr-Value test is run to
find the candidate’s percentages of gaining a seat in the Brazilian Chamber of Deputies
based on their cash received.
The last test that is used deals with summary statistics of the candidates from the
Sao Paolo district. Two different sets of summaries are established. First, all the
candidates with a quality rating of three are put into one dataset. Then, summary
statistics are performed on these candidates to pinpoint their average funding received
and percentage of votes received. Summary statistics are also found on the second
dataset, which includes all of the candidates from the Sao Paolo district with quality
ratings of one. I feel that these statistics from both datasets can be compared and
analyzed with the percentages from the “Pr-value” test to show that in extreme cases
where high quality candidates are facing low quality candidates, money will tend to
matter more to the low quality group.
Finally, I plugged the numbers from the previous summary statistics of the
average funding received back into the Prvalue command to determine the percent chance
14. 14
on average that Incumbents with a quality rating of three and challengers with a quality
rating of one have to gain a seat in the Brazilian Chamber of Deputies. All the results are
displayed below.
Results
The first logistic regression test on the overall data shows that both the percent
cash and the incumbent variable are extremely significant to the dependent variable;
gaining a seat or not in the Brazilian Chamber of Deputies. The one difference included
in the data received from Professor David Samuels was the yes_no variable, which
described whether or not a candidate gained a seat. This variable essentially became the
dependent variable in the statistical analysis. The table with the statistics of the logistic
regression test is listed below.
[Insert Table 1)
The next test dealt with summary statistics when isolating the candidate quality
variable. The results of the summary statistics are very intriguing. The statistics for
candidates with a quality rating of three showed that on average, these candidates gained
a seat in the Brazilian Chamber of Deputies more often, they received more of the
percentage cash, and they received more of the percentage vote in their perspective
districts than candidates with a rating of one. The main variable that the research looks at
is the percentage cash received. Candidates with a quality of three on average received
3.35% of the campaign funds, while candidates with a quality rating of one on average
15. 15
received 2.48 % of the cash. Obviously this shows how important being a quality
candidate is for election purposes. The table is shown below.
[Insert Table 2]
The next logistic regression results were also very intriguing. At first, a logistic
regression test was performed on all of the data provided by Professor Samuels. For this
test, I ran the data while filtering candidates who didn’t have an extreme quality rating.
So only challengers with ratings of one and incumbents with ratings of three were
included. What was interesting about these results was the fact that the test dropped the
candidate quality variable because of colinearity with the dependent variable. In other
words, the candidate quality variable almost perfectly explained whether or not the
incumbent or challenger gained a seat. This did not occur in the first logistic regression
test, which is important because it shows that when isolating the variable to only extreme
candidates, the candidate quality variable becomes extremely significant. The table is
listed below.
[Insert Table 3]
After the last logistic regression test was completed, the “Pr-value” command was
then used to find out the percentage chance that an incumbent and challenger have to gain
a seat in the Brazilian Chamber of Deputies. I ran this test on the largest district in
Brazil, with a magnitude of seventy. The results showed that when a candidate gained
16. 16
5.4% of the funds in the district, they had a 100% chance of gaining a seat. When a
candidate gained .99% of the funds, they had a 90% chance of gaining a seat. The
candidates had an 80% chance of gaining a seat when they received .526% of the funds,
and they had a 70% chance of gaining a seat when they received .218% of the funds. The
table is listed below.
[Insert Table 4]
The last of the results includes more analysis on the largest district in Brazil, Sao
Paolo (Magnitude 70). The results show summary statistics of the candidates, while
again isolating the extreme quality incumbents and challengers. In the Sao Paolo district
for the 1989 election, thirty-two candidates received a quality rating of three, while
twenty-three received a rating of one. Eighteen of the candidates who were given a rating
of three were incumbents while only fourteen were challengers. What is striking is the
fact that only one incumbent in the district was given a quality rating of one, while the
other twenty-two candidates given a one were challengers. So it can obviously be said
that challengers in this district are at a heavy disadvantage in quality ratings. Another
significant result showed that out of the thirty-two candidates with a rating of three,
100% gained a seat. Out of the twenty-three candidates given a rating of one, only 47.8%
gained a seat.
Also in this table, percentages are shown that explain the chance the candidates
had to gain a seat based on the funds they received. The average amounts of cash
received for incumbents with a rating of three and challengers with a rating of one were
17. 17
plugged into a “Pr-value” command. The results showed that on average, candidates
with a quality rating of three had an 86.7% chance of gaining a seat, while candidates
with a rating of one had a 75.2% chance. The results seem to show that candidate quality
has a great impact on electoral success. The table is shown below.
[Table 5]
Conclusion
As the tests have shown, candidate quality seems to have a significant impact on
electoral success in the Brazilian Chamber of Deputies. As stated before, many scholars
have attempted to explain the importance of campaign spending to candidates in any
legislative system. This paper has attempted to review and analyze the research of David
Samuels on campaign spending in the Brazilian Chamber of Deputies. However, I have
focused in on the candidate quality variable with great detail. The research attempts to
challenge the notion that incumbents and challenegers running for a seat in the Brazilian
Chamber of Deputies value campaign spending equally due to a weak incumbency
advantage present there. Instead, I argue that when extreme candidates, based on
candidate quality ratings, are competing against each other, the lower quality candidate
will value campaign spending more.
The results of all of the tests give sufficient evidence showing that candidate
quality has a definite impact on electoral success. The research also shows that when
isolating the candidate quality variable, and testing high quality incumbents against low
quality challengers in their perspective districts, money will tend to matter more to the
challenger. Through the research, I feel I can confidently argue that when extreme
18. 18
candidates, based on candidate quality, are competing against one another, campaign
spending will matter more to the lower quality candidate.
Also through this research, a major question has definitely been left unanswered.
How does candidate quality affect the importance of campaign spending in other
countries? Interest in this question has been portrayed throughout this paper. Similar to
how I challenge the claim towards the Brazilian Chamber of Deputies, future research
can be conducted to study the United States House of Representatives. Do challengers
always value campaign funding and spending more than incumbents do, as previous
research has shown? Is there really always a problem of diminishing returns for
incumbents? I think, based on this research, that a future theory can be presented arguing
that when high quality challengers compete against low quality incumbents, campaign
funding will be equally if not more important to the incumbent in the United States.
19. 19
References
Ames, Barry. (2001). The Deadlock of Democracy in Brazil. University of
Michigan Press, 331.
Coleman, J. J., & Manna, P. F. (2000). Congressional Campaign Spending and the
Quality of Democracy. Journal of Politics, 62(3), 757-789.
Green, D. P., & Krasno, J. S. (1988). Salvation for the Spendthrift Incumbent:
Reestimating the Effects of Campaign Spending in House elections.
American Journal of Political Science, 884-907.
Jacobson, G. C. (1978). The Effects of Campaign Spending in Congressional
Elections. The American political science review, 469-491.
Jones, M.P. (2002). Legislative Behavior and Executive-Legislative Relations in
Latin America. Latin American Research Review, 37(3), 176-188.
Samuels, D. (2001). Does Money Matter? Credible commitments and campaign
finance in new democracies: theory and evidence from Brazil. Comparative
Politics, 23-42.
Samuels, D. (2001). Incumbents and Challengers On a Level Playing Field:
Assessing the impact of campaign finance in Brazil. The journal of Politics,
63(02), 569-584.
Samuels, D. (2001). Money, Elections, and Democracy in Brazil. Latin American
Politics and Society, 43(2), 27-48.
Samuels, D. (2004). Presidentialism and Accountability for the Economy in
Comparative Perspective. American Political Science Review, 98, 425-436.
20. 20
Table
1
Logistic
Regression:
David
Samuels’
Campaign
Spending
Data
Including
the
yes_no
Variable
Logistic
regression
Number
of
obs
=
592
LR
chi2(2)
=
51.82
Prob
>
chi2
=
0.0000
Log
likelihood
=
-‐294.35377
Pseudo
R2
=
0.0809
-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐
yes0_no1
|
Coef.
Std.
Err.
z
P>|z|
[95%
Conf.
Interval]
-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐+-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐
cash
|
-‐.3015248
.0627123
-‐4.81
***0.000
-‐.4244387
-‐.1786109
incumbent
|
-‐.5302329
.2127577
-‐2.49
***0.013
-‐.9472302
-‐.1132355
_cons
|
-‐.3967193
.1541656
-‐2.57
0.010
-‐.6988783
-‐.0945604
-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐
21. 21
Table
2
Variable
|
Obs
Mean
Std.
Dev.
Min
Max
-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐+-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐
yes0_no1
|
188
.0585106
.2353332
0
1
cash
|
188
3.354705
3.93233
.0336787
22.28703
vote
|
188
1.899043
1.814032
.0228288
9.731059
Variable
|
Obs
Mean
Std.
Dev.
Min
Max
-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐+-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐-‐
yes0_no1
|138
.4710145
.5009776
0
1
cash|
138
2.482561
4.378081
.0000204
26.04973
vote|
138
1.113453
1.319865
.0187311
8.541571
SUMMARY
STATISTICS
FOR
CANDIDATES
WITH
A
QUALITY
RATING
OF
1
SUMMARY
STATISTICS
FOR
CANDIDATES
WITH
A
QUALITY
RATING
OF
3
23. 23
Table
4
PERCENTAGE
OF
FUNDS
RECEIVED
Pr(y=0/x)
Chance
of
Gaining
a
Seat
Pr(y=1/x)
Chance
Not
Gaining
a
Seat
5.4%
100%
0%
.99%
90%
10%
.526%
80%
20%
.218%
70%
30%
Threshold
to
Win
Seat
in
District
With
Magnitude
of
70
24. 24
Table
5
Candidates
with
a
Quality
Rating
of
3
Candidates
with
a
Quality
Rating
of
1
Thirty-‐Two
Candidates
received
a
Quality
Rating
of
3
Twenty-‐Three
Candidates
received
a
Quality
Rating
18
Incumbents
14
Challengers
One
Incumbent
Twenty-‐Two
Challengers
32
Out
of
32
Gained
a
Seat
100%
11
Out
of
23
Gained
a
Seat
47.8%
Average
Percent
of
Vote:
.495%
Average
Percent
of
Cash
Received:
.807%
Average
Percent
of
Vote:
.155%
Average
Percent
of
Cash
Received:
.368%
Analysis
of
Sao
Paolo
District:
Magnitude
70
*When
Using
PR-‐value
test,
Candidates
with
a
Quality
Rating
of
3
have
a
higher
chance
of
gaining
a
seat.
CANDIDATE
(QUALITY
3):
86.7%
CANDIDATE
(QUALITY
1):
75.2%