Policy Uncertainty, Political Capital, and Firm Risk-Taking∗
Pat Akey†
University of Toronto
Stefan Lewellen‡
London Business School
November 3, 2015
Abstract
We link the cross-section of firms’ sensitivities to economic policy uncertainty to their subse-
quent political activity and post-election risk-taking behavior. We first show that firms with a
high sensitivity to economic policy uncertainty donate more to candidates for elected office than
less-sensitive firms. Using close election outcomes for identification, we then compare differences
in post-election risk-taking and performance for firms that experience the same close-election
political capital shocks and the same general policy uncertainty shock but that differ in their
ex-ante sensitivities to policy uncertainty. We find that policy-sensitive firms’ investment, lever-
age, operating performance, Tobin’s Q, option-implied volatility, and CDS spreads respond more
sharply to the resolution of uncertainty than policy-neutral firms. However, these effects are
largely restricted to firms experiencing a negative political capital shock, suggesting that po-
litical connections are an imperfect hedge against shocks to the economic policy environment.
Our results highlight many new links between political capital shocks and firms’ subsequent
decision-making and represent the first attempt in the literature to link firms’ policy uncer-
tainty sensitivities to their political activities and subsequent risk-taking and performance.
∗The authors would like to thank Ling Cen, Olivier Dessaint, Art Durnev, Nicholas Hirschey, Chay Ornthanalai,Mike Simutin, Vania Stavrakeva, and Pietro Veronesi for valuable comments.†University of Toronto. Phone: +1 (647) 545-7800, Email: [email protected]‡London Business School. Phone: +44 (0)20 7000 8284. Email: [email protected]
1
1 Introduction
Government policy represents a large source of uncertainty for many corporations, yet there is
currently little research on the question of how economic policy uncertainty affects firms’ operating
decisions and performance. In this paper, we use firms’ political activities as a lens through which
to study the relationship between economic policy uncertainty and firms’ risk-taking behavior.
We do so by first linking economic policy uncertainty to firms’ donations to candidates in U.S.
congressional elections. We then examine how the resolution of uncertainty following election
outcomes affects firms’ subsequent risk-taking and performance, and in particular, how risk-taking
and performance differ between (ex-ante) policy-sensitive and policy-neutral firms following shocks
to these firms’ political capital bases stemming from close election outcomes.
Our focus on political activities is driven by the observation that, much as firms can hedge
against other forms of uncertainty such as price uncertainty or weather uncertainty, firms can also
“hedge” against economic policy uncertainty by actively attempting to influence the policy-making
process. Moreover, firms that are highly sensitive to changes in government policy arguably have a
stronger incentive to engage in the political process than similar, policy-neutral firms.1 Hence, in
order to accurately measure the effects of economic policy uncertainty on firms’ subsequent risk-
taking, we must also account for firms’ attempts to influence the policy-making process through
the accumulation and deployment of political capital. This intuition – which to our knowledge is
new to the literature – also suggests that firms may participate in the political process due to policy
uncertainty considerations alone, which represents a sharp contrast to the “increased government
handouts” and “increased credit availability” political participation channels that are popular in
the literature on political connections.
The most basic form of political capital is a direct connection between firms and policymakers.
As such, we focus on firms’ campaign contributions to candidates for elected office as our measure
1While there are no theories (to our knowledge) that link together uncertainty, political connections, and firmrisk-taking, we can appeal to the literature on hedging to support this argument (see, e.g., Holthausen (1979)). Inthe presence of financing frictions, taxes, bankruptcy costs, or other types of frictions, a positive shock to uncertaintywill increase the demand for hedging holding the firm’s production function constant. All else equal, this implies thatthe marginal value of an extra hedging unit will be larger for firms exposed to greater levels of uncertainty.
2
of political capital. We identify a subset of firms that are particularly sensitive to economic policy
uncertainty within a given U.S. congressional election cycle and then track these firms’ subsequent
donation activities. Consistent with the intuition that the marginal value of a political connection
is greater for policy-sensitive firms, we find that policy-sensitive firms are more likely to make
(or increase) political campaign contributions relative to policy-neutral firms. Using close election
outcomes as a plausibly exogenous shock to firms’ political capital bases, we then examine how
shocks to a firm’s political capital stock impact the firm’s subsequent risk-taking behavior. In
particular, by examining differences in subsequent risk-taking across policy-sensitive and policy-
neutral firms that are subject to the same political capital shocks, we can directly isolate the effect
of economic policy uncertainty on risk-taking after controlling for firms’ own abilities to influence
the policy-making process. Furthermore, by comparing outcomes across policy-sensitive (or policy-
neutral) firms that experience different political capital shocks, we are able to measure the marginal
effects of an extra political connection on firm risk-taking and performance and are able to isolate
differences in these effects across policy-sensitive and policy-neutral firms. This latter set of tests
is important because it allows us to link our findings on the marginal value of political connections
to the existing political connections literature.
Our empirical design allows us to overcome many of the difficulties associated with identifying
the relationship between policy uncertainty and firm outcomes. First, as noted previously, our focus
on firms’ political donations allows us to directly control for firms’ “hedging” activities, thereby
allowing us to compare policy-sensitive (policy-neutral) firms that experience similar political cap-
ital shocks. Second, our sample of close congressional elections in the United States allows us to
identify true “shocks” to donating firms’ political capital stocks (since by definition, close elections
are not decided until election day).2 By limiting our focus to the sample of firms that actively
donated money to candidates in close elections, we are also able to mitigate concerns related to the
endogeneity of firms’ decisions to participate in the political process. Furthermore, we use ex-ante
measures of policy uncertainty sensitivity to avoid the simultaneity problems caused by firms’ joint
2Our primary identifying assumptions are that close election outcomes are unknown at the times of firms’ dona-tions, and that firms’ donations themselves are not sufficient to sway election results. While neither assumption isdirectly testable, we present evidence later in the paper that is consistent with both assumptions.
3
management of their policy uncertainty sensitivities and their political capital activities. Finally,
most of our specifications include firm-election cycle fixed effects (that is, a separate fixed effect ex-
ists for each firm during each election cycle). This allows us to control for any unobserved variation
within each firm-election cycle pairing that may be driving our results.3
To estimate the relationship between economic policy uncertainty and risk-taking, we begin
by sorting firms into “policy-sensitive” and “policy-neutral” buckets six months before each U.S.
federal election using data from the previous 18 months. We identify firms as being policy-sensitive
based on a regression of firm stock returns on the Economic Policy Uncertainty index developed by
Baker, Bloom, and Davis (2015) and various control variables. Within each election cycle, we then
define the magnitude of the ex-post political capital shock for firm i as the difference between the
number of ultimate winners and losers that the firm supported in close elections during that cycle.4
We define a firm as receiving a “lucky” political capital shock if, in a given election cycle, the
firm’s close-election candidates win more elections than they lose.5 Consistent with our primary
identifying assumptions, the median political capital shock in our sample is exactly zero. The
risk-taking and performance variables that we examine include market-driven variables (such as
option-implied volatility, CDS spreads, and Tobin’s Q) and firm-driven investment and operating
variables (such as investment, leverage, R&D spending, operating profitability, and sales growth).6
We then employ a differences-in-differences design with multiple treatment groups to examine
how “lucky” and “unlucky” shocks to firms’ political capital bases affect their subsequent risk-taking
behavior.7 We also use triple-difference specifications to examine whether the effects we observe for
3Our results are also robust to the use of industry-cycle fixed effects.4For example, Coca-Cola donated to two winning candidates and five losing candidates in close elections during
the 2004 election cycle, so we compute the shock to Coke’s political capital during the 2004 cycle as 2 - 5 = -3. Incontrast, Coke supported seven close-election winners and four close-election losers during the 2006 election cycle, soCoke’s political capital shock in the 2006 cycle would be defined as 7 - 4 = 3.
5This approach implicitly assigns an identical weight to every election. In reality, some elections may be moreimportant than others. However, there is no reason to believe that this approach induces a systematic bias in ourmeasurement of political capital shocks, which would be required in order for this concern to affect our results. Asa further robustness check, we also limit our sample to firms that make donations prior to September 1st within agiven election cycle (federal elections take place in the United States on the first Tuesday of November). All of ourresults remain quantitatively and qualitatively unchanged.
6These variables have been used before (in different contexts) within the existing literature on political connections.7Standard difference-in-difference designs contain a treatment group and a control group. Here, both groups are
treated: one experiences a positive shock while the other experiences a negative shock. This experimental design isarguably better suited than standard difference-in-difference designs for causal inference, since, in the words of Cookand Campbell (1979), “the theoretical causal variable has to be rigorously specified if a test is made that depends on
4
“lucky” winners versus “unlucky” losers differ between policy-sensitive and policy-neutral firms.
Hence, in total, we are able to isolate the effects of political capital shocks on firm risk-taking
for four different types of firms: “lucky” policy-sensitive firms, “unlucky” policy-sensitive firms,
“lucky” policy-neutral firms, and “unlucky” policy-neutral firms. This decomposition allows us to
directly test the relationships between policy uncertainty, political capital, and firms’ subsequent
risk-taking and performance.
Our analysis yields four main results, all of which are new to the literature. First, holding
policy sensitivity fixed, we find that “lucky” political capital shocks are associated with a subsequent
reduction in firms’ risk-taking (as measured by variables such as implied volatility and CDS spreads)
and an improvement in firms’ operating performance (as measured by variables such as Sales and
ROA). We also find that “lucky” political capital shocks are associated with increases in firm
value (as measured by Tobin’s Q). These effects are opposite in sign but are roughly symmetric in
magnitude for “lucky” versus “unlucky” firms, supporting our identifying assumption that election
outcomes were unknown at the time of firms’ campaign contributions in our sample of close elections.
We next compare the magnitudes of the differences between “lucky” and “unlucky” shocks
across firms with different ex-ante policy sensitivities. Consistent with the marginal value of a
political connection being larger for policy-sensitive firms, we find that the differences we observe in
post-election outcomes across “lucky” and “unlucky” policy-sensitive firms are larger in magnitude
than the differences we observe across “lucky” and “unlucky” policy-neutral firms. The economic
magnitudes of these differences are significant: for example, we observe a 10% relative difference
in investment levels, a 2% relative difference in leverage, a 13% relative difference in Tobin’s Q,
a 12% relative difference in one-month option-implied volatilities, and a 10% relative difference
in one-year CDS spreads. These findings confirm our intuition that policy-sensitive firms respond
more sharply to political capital shocks relative to policy-neutral firms.
To estimate the effects of policy uncertainty sensitivity on risk-taking, we examine the differ-
one version of the cause affecting one group one way and another group the [opposite] way...[i]t is the high constructvalidity of the cause which makes [this] design potentially more appropriate for theory-testing research than the[standard difference-in-difference] design.” Empirically, the recovered treatment effects from our design are identicalto the treatment effects that would arise from a standard (and properly-specified) difference-in-difference analysis.
5
ences in treatment responses across policy-sensitive and policy-neutral firms experiencing the same
political capital shock. In other words, we examine differences in outcomes between “lucky” policy-
sensitive firms and “lucky” policy-neutral firms (and vice versa). These firms differ in their ex-ante
policy sensitivities, but both firms experience similar political capital shocks. As such, we are able
to estimate the differential response of policy-sensitive versus policy-neutral firms to changes in
policy uncertainty after controlling for the effects of political capital shocks.
Interestingly, we find that the resolution of policy uncertainty has an asymmetric impact on
firms’ risk-taking and performance depending on whether firms are hit with a lucky or unlucky
political capital shock. Policy-sensitive firms that are hit with an unlucky political capital shock
have lower investment, higher leverage, lower Q, worse operating performance, and higher implied
volatility and CDS spreads than policy-neutral firms hit with a similarly unlucky shock. Hence,
consistent with intuition, we find that unlucky political capital shocks hurt policy-sensitive firms
particularly badly relative to their policy-neutral brethren. However, even when hit with a positive
political capital shock, we find that policy-sensitive firms invest less, have higher CDS spreads,
and have higher implied volatilities than policy-neutral firms hit with similarly positive shocks
(though the magnitudes are significantly smaller). Hence, we find that the effects of policy uncer-
tainty on risk-taking and performance are asymmetric: firms hit with a bad political capital shock
suffer greatly, while firms hit with a good shock are still negatively impacted, though to a lesser
degree. These findings suggest that political connections serve as an imperfect hedge against policy
uncertainty, since even the “gains” associated with improved political connections do not allow
policy-sensitive firms to completely offset the effects of being sensitive to future economic policy
changes. Nevertheless, the differences in the magnitudes of the treatment responses across policy-
sensitive and policy-neutral firms still suggest that on the margin, an extra political connection is
still more valuable for policy-sensitive firms than for policy-neutral firms.
Finally, we examine the term structure of political capital shocks using data on long-dated
options and CDS contracts. We find that political capital shocks seem to have long-lasting effects
on firms’ (expected) risk-taking: 6-month implied volatilities and 10-year CDS spreads drop signif-
6
icantly following elections for “lucky” firms, and the magnitudes of the drops are consistent with a
multi-year shock to firms’ risk-taking. We also find that implied volatility slopes and CDS slopes
steepen modestly for “lucky” firms following close elections, which suggests that the “half-life” of
the effects of political capital shocks on firm riskiness is potentially long. These results are largely
consistent with our other findings reported above.
Our results also point to a “policy uncertainty” channel of political capital accumulation that is
distinct from the channels that have been previously documented in the literature. For example, a
growing literature points to firms’ abilities to secure government funds through “bailouts” (Faccio,
Masulis, and McConnell (2006); Duchin and Sosyura (2012)) or through other forms of government
spending (Cohen and Malloy (2014)) as the primary channel through which firms benefit from
political connections. While we cannot rule out this channel, most of our results suggest that
risk-taking declines and operational performance improves following a positive political capital
shock, which is not consistent with the existing literature on the “increased government handouts”
channel.8 Another channel argues that politically-connected firms benefit from increased credit
availability (often through loans made by politically-connected banks), potentially reducing these
firms’ financial constraints (see, e.g., Khwaja and Mian (2005) and Claessens, Feijen, and Laeven
(2008)). However, our findings suggest that firms’ average leverage decreases following positive
political capital shocks, which is not consistent with the “increased credit availability” story. In
short, while the ultimate motives behind firms’ political donation decisions are unobservable, our
results line up most closely with the idea that corporations cultivate political connections in an
attempt to influence policy implementation on issues that are particularly relevant to the firm.
2 Related Literature
A recent strand of literature examines the effects of aggregate policy or political uncertainty on firm
outcomes and asset prices. Pastor and Veronesi (2012, 2013) model how asset prices and risk premia
respond during times of high aggregate political uncertainty. Kelly, Pastor, and Veronesi (2014)
8We also test the “sticky-government-contracting” hypothesis explicitly and find no evidence that government-dependent firms are driving our results.
7
find that political uncertainty is priced in options markets. Julio and Yook (2012) and Durnev
(2010) find that investment levels and stock price-investment sensitivities decrease during periods
of high political uncertainty, while Boutchkova, Doshi, Durnev, and Molchanov (2012) document
that export-oriented industries, industries dependent on contract enforcement, and labor-intensive
industries exhibit higher volatility during periods of political uncertainty. Finally, Gulen and Ion
(2015) find that investment and policy uncertainty are negatively related, particularly for firms that
have high investment irreversibility and dependence on government revenues. To date, however,
this strand of the literature has not looked at how economic policy uncertainty interacts with firms’
political connections.
Our paper is also related to two strands of the literature on political connections. One strand
focuses on the link between political connections and firm value. In particular, Fisman (2001),
Faccio (2004), Faccio and Parsley (2009), Cooper, Gulen, and Ovtchinnikov (2009), Goldman,
Rocholl, and So (2009), Acemoglu, Johnson, Kermani, Kwak, and Mitton (2013), Akey (2015),
and Borisov, Goldman, and Gupta (2015) all find evidence that stronger political connections are
associated with increases in firm value.9 Consistent with this literature, we find that unexpected
positive shocks to firms’ political capital stocks are associated with increases in firm value (as
measured by Tobin’s Q).
A second strand of the political connections literature is focused on identifying the underlying
rationale for firms to establish connections with politicians in the first place. Faccio, Masulis,
and McConnell (2006) and Duchin and Sosyura (2012) show that politically-connected firms are
more likely to receive government bailouts than non-connected firms, while Cohen and Malloy
(2014) find that government-dependent firms (who are likely to be politically-connected) have
lower investment, lower R&D spending, and lower sales growth than non-government-dependent.
Relatedly, Kim (2015) finds that firms with strong political connections have lower investment,
lower R&D spending, and lower patent citations (but higher government sales) relative to firms with
weak political connections. These findings are largely consistent with the theoretical predictions of
9In contrast, Agarwal, Meshke, and Wang (2012) and Coates IV (2012) find that political connections mayindicate agency problems in connected firms. However, the overwhelming majority of studies has found that politicalconnections have a large and positive impact on firm value.
8
Murphy, Shleifer, and Vishny (1993) and Shleifer and Vishny (1994).10 In contrast, Khwaja and
Mian (2005) and Claessens, Feijen, and Laeven (2008) find that politically-connected firms benefit
from increased credit availability and a potential reduction in financial constraints. In addition,
Do, Lee, and Nguyen (2013) finds that firms invest more in physical capital, while Hanouna,
Ovtchinnikov, and Prabhat (2014) find that CDS spreads on average tend to be lower for politically-
connected firms. Finally, a number of studies have found that political connections lead to higher
future sales (Amore and Bennedsen (2013), Goldman, Rocholl, and So (2013), Tahoun (2014),
Akey (2015)). However, none of these papers examines the effects of policy uncertainty on political
capital and firms’ subsequent risk-taking.
There is one other paper in the literature (Ovtchinnikov, Reza, and Wu (2014)) that links pol-
icy uncertainty to political connections and political connections to firm outcomes (in this case,
innovation). In particular, Ovtchinnikov, Reza, and Wu (2014) find that firms’ innovation increases
following positive political capital shocks, which they argue is due to a reduction in policy uncer-
tainty amongst politically-connected firms. Our paper differs from Ovtchinnikov, Reza, and Wu
(2014) along many dimensions: we use a different identification strategy, our sample consists of
a different set of political connections, and we measure different outcome variables. Furthermore,
Ovtchinnikov, Reza, and Wu (2014) do not explicitly test whether policy uncertainty is directly
impacting political connections or risk-taking. Nevertheless, the key message of their paper – that
politically-connected firms may benefit from a reduction in policy uncertainty – nicely complements
the main findings of our paper.
3 Economic Setting and Identification Strategy
3.1 Economic Setting and Testable Implications
Our goal is to study the interaction between firms’ sensitivity to economic policy uncertainty, their
subsequent political activities, and their ultimate response to the resolution of policy uncertainty
10Johnson and Mitton (2003) find that politically-connected firms benefit from foreign capital controls and sufferwhen these controls are removed, consistent with the predictions of Rajan and Zingales (1998). These findings arealso in the spirit of Murphy, Shleifer, and Vishny (1993) and Shleifer and Vishny (1994).
9
following U.S. federal elections. Our basic argument consists of four main components. First,
some firms are more exposed (or sensitive) to economic policy uncertainty than other firms. For
example, financial institutions and auto manufacturers may have been more exposed to policy
uncertainty ahead of the 2008 elections than, say, textile producers. We are agnostic as to the
source of firms’ exposure to policy uncertainty – for example, firms’ exposure may stem from a
“shock” to the policy environment or may reflect firms’ own decision-making. We simply wish to
argue (and document) that prior to each U.S. federal election in our sample, some firms are more
exposed to policy uncertainty than others. We label firms with high exposure to policy uncertainty
as “policy-sensitive firms.”
We next conjecture that policy-sensitive firms are more likely to make (or increase) political
campaign contributions than less-policy sensitive firms (which we refer to as “policy-neutral” firms).
In other words, to the extent that political connections have value to firms, we posit that the
marginal value of a political connection is higher for policy-sensitive firms than for policy-neutral
firms. As noted in the introduction, this hypothesis is consistent with the literature on firms’
hedging behavior under uncertainty in the presence of market frictions. Our hypothesis leads
immediately to our first testable prediction, which is that policy-sensitive firms will contribute
more to candidates for elected office than policy-neutral firms within a given election cycle.
Having linked policy uncertainty sensitivity to firms’ campaign donations, we next link firms’
donation choices to the outcomes of U.S. federal elections. Economic policy uncertainty tends to
rise before U.S. federal elections and declines following these elections (Baker, Bloom, and Davis
(2015)). One interpretation of this finding is that a significant proportion of existing (ex-ante)
policy uncertainty is related to a given set of elections and is resolved by the outcomes from these
elections. Under this interpretation, the uncertainty faced by policy-sensitive firms should decline
following elections, and should decline more so than for non-policy sensitive firms. As such, we
posit that the previously-documented tendency for firms to “wait” on the outcomes of elections
(Julio and Yook (2012), Pastor and Veronesi (2013), Kelly, Pastor, and Veronesi (2014)) will lead
to sharper post-election changes in firm operating behavior for policy-sensitive firms relative to
10
non-policy sensitive firms. This is another prediction that is directly testable in our sample.
Election outcomes resolve two types of uncertainty: uncertainty related to future government
policies, and uncertainty regarding a firm’s stock of political connections or political capital. As
such, firms’ responses to the resolution of policy uncertainty may depend in part on whether the
firm’s own stock of political capital has been strengthened or weakened. Under the assumption that
firms’ campaign contributions are linked to its policy objectives, this implies that firms experiencing
a “lucky” election draw (i.e. an election where many of the firms’ contributions went to victori-
ous candidates) will respond differently to the resolution of political uncertainty relative to firms
whose contributions primarily went to losing candidates. For example, “winning” firms may decide
to increase investment, while “losing” firms may decide to postpone investment.11 Furthermore,
policy-sensitive firms should be expected to increase or decrease investment (or other variables)
even more than their non-policy sensitive brethren. These two predictions are also directly testable
and arguably represent the main contribution of this paper.
One last question remains, however, and that is how firms should respond to the resolution of
uncertainty given a favorable or unfavorable election draw. The answer to this question depends in
part on the goals that firms have when they make political contributions to candidates for elected
office. For example, if a firm establishes political connections for the purpose of securing future
“bailouts” or other types of hedges against bad economic states of nature, this implies that a positive
shock to a firm’s political capital base may be followed by an increase in risk-taking behavior.
Alternatively, if a firm establishes political connections for the purpose of securing government
contracts, a positive political capital shock might be followed by a reduction in risk-taking behavior,
since extra risk could plausibly put the firm in jeopardy of securing future contractual agreements
with the government (or, as another possibility, firms may simply kick back and enjoy the “quiet
life”). Finally, if a firm establishes political connections in order to simply influence the policy-
making process (perhaps because the firm is policy-sensitive), then a positive shock to political
11The expected signs of these effects are theoretically ambiguous. For example, moral hazard arguments suggestthat stronger political connections should be linked to an increase in firm risk-taking. In contrast, governmentcontracting considerations may cause firms to decrease risk-taking following positive political capital shocks becausegovernment funding may crowd out the firms’ own investment activities.
11
capital is likely to improve the probability of the firm obtaining its policy objectives, which in turn
may lead the firm to either increase or decrease its subsequent risk-taking behavior depending on
the firm’s specific policy objectives.
We formally test each of these three “channels” – along with the broader links between policy
uncertainty, campaign contributions, election outcomes, and subsequent risk-taking – using an iden-
tification strategy that exploits plausibly exogenous variation in close U.S. congressional elections.
We now describe this approach.
3.2 Identification and Empirical Approach
Estimating the causal effect of political capital shocks on ex-post firm outcomes is extremely chal-
lenging. First, firms endogenously choose whether to be politically active and which politicians
to form connections with. Second, certain types of firms may be more likely to donate to certain
types of candidates who are themselves more or less likely to be elected (for example, powerful
incumbents). Third, the results of most elections are effectively determined months before the
actual election date, making it difficult to isolate the timing of any political-capital-related shocks
on market prices. Fourth, the causality could go in the other direction; that is, firms’ operating
decisions or riskiness may affect the outcome of elections and/or create shocks to the firm’s political
capital ledger.12 Finally, other sources of unobserved heterogeneity may account for any observed
relationship between political capital shocks and firms’ riskiness and operating decisions. For ex-
ample, a disruptive technology shock may jointly affect firms’ operating decisions as well as the
outcome of political elections in the state(s) most affected by the change.
To overcome these challenges, we focus on a subset of firms that donate to candidates in “close”
U.S. congressional elections from 1998-2010.13 Our primary identifying assumption is that election
outcomes at the time of firms’ donations are plausibly exogenous in our sample of close elections.
Our claim of plausible exogeneity requires two key assumptions to be met: first, firms cannot
12For example, financial institutions’ behavior prior to the recent crisis may have affected the outcome of electionsand/or the firms’ political capital.
13Following Akey (2015), Do et al. (2012), and Do et al. (2013), we define a close election as an election that isultimately decided by a voting margin of 5% or less. In other words, for an election with two candidates, we definea close election as one in which the winner receives 50.01% - 52.5% of the total vote.
12
accurately predict close-election winners at the time of their donations, and second, firms’ donations
themselves cannot materially affect a candidate’s chances to win an election. While neither of
these assumptions are directly testable, anecdotal evidence strongly supports the view that election
outcomes in our sample are plausibly random conditional on firms’ donation decisions. In particular,
we find that the median firm in our sample supports exactly the same number of close-election
losers as close-election winners during each congressional election cycle. Consistent with this fact,
Figure 1 shows that the distribution of firms’ net close-election political capital shocks is centered
around zero, is effectively unimodal, and has relatively symmetric tails. Under the assumption that
firms would rather donate to winning candidates than losing candidates, these results suggest that
election outcomes are largely unpredictable in our sample of close elections and that a given firm’s
donations are not sufficient to sway election outcomes. Furthermore, by looking within the set of
firms that made close-election donations, we are able to effectively control for the fact that donation
patterns are not random, since all of the firms in our sample felt that it was optimal (for whatever
reason) to donate to one or more close-election candidates. Finally, close elections are generally
decided on election day (or very soon before), making it easier to isolate the timing associated with
a market response to political capital shocks.
Our identification strategy allows us to examine outcome differences between firms who donated
to a politician that just won a close election relative to firms who donated to a politician that just
lost a close election in the time periods before and after each election. Our implicit assumption is
that these “random” election-day outcomes represent positive or negative shocks to the political
capital of donating firms. If political capital shocks have an effect on firms’ riskiness or operating
decisions, we should see firms (and markets) respond to election-day political capital shocks in the
ensuing days, weeks, and months following U.S. federal elections. In the context of the economic
setting that was discussed in the previous section, these shocks can be interpreted as an increase
or a decrease in connectedness. This hypothesis forms the basis for our empirical strategy, which
we describe below.
We begin by defining Close Winsi,t (Close Lossesi,t) as the number of close-election winners
13
(losers) that firm i donated to during election cycle t. For example, since Coca-Cola donated to
two close-election winners and five close-election losers during the 2004 election cycle, we would
set Close Wins = 2 and Close Losses = 5 for Coke during the 2004 cycle. We next define
Net Close Winsi,t as the difference between Close Wins and Close Losses. This variable captures
a firm’s overall political capital gain in close elections during a given cycle. For Coke in 2004, this
variable would be defined as Net Close Wins = 2 − 5 = −3. We also create a dummy variable
(Close Election Dummy) that takes the value of one if firm i’s overall political capital gains are
greater than the sample median of zero (i.e. where Net Close Winsi,t > 0) during a given election
cycle, and takes the value of zero otherwise. For example, since Coke donated to more close-election
losers than winners in 2004, we would set Close Election Dummy equal to zero for Coke in 2004.
Importantly, our close-election measures are not strongly correlated with general election out-
comes: for example, in the 2006, 2008, and 2010 elections (which were “dominated” by Democrats,
Democrats, and Republicans, respectively), our Close Wins variable suggest that more of the close
winners in 2006 and 2008 were Republicans, while there is no clear pattern in 2010. As such, our
close election results do not appear to be picking up “general” election trends. This helps to rule
out the possibility that our close-election results are simply picking up “general” election shocks
that tend to benefit firms that donated (nearly) exclusively to the party that “won” the election,
perhaps because of similar ideological positions on government policy priorities.
The variables Close Wins, Close Losses, Net Close Wins, and Close Election Dummy form
our primary measures of political capital shocks. We will use all four variables in our subsequent
tests. Panel A of Table 1 presents summary statistics for each of these measures (with the exception
of Close Election Dummy, which is an explicit function of the sample median).
With our political capital measures in hand, we can now turn to our empirical framework. We
employ a differences-in-differences framework to estimate the effects of a political capital shock on
14
firm outcomes.14 Specifically, we estimate the following model:
Outcomei,t =α+ β1PostElectiont + β2Capital Shocki,t × PostElectiont+ (1)
+ Γ′Controlsi,t + Firm× ElectionCycle FE + εi,t ,
where Capital Shocki,t represents one of our political capital shock measures described above.
Unlike standard differences-in-differences tests, there is no separate term for Capital Shocki,t in
the equation above because the granularity of our data allows us to include firm-election cycle fixed
effects (Capital Shocki,t is defined at the firm-cycle level, so it is swept up by our fixed effects). As
such, our results can be interpreted as looking within a firm and given election cycle.
Our primary coefficient of interest is β2 in the equation above. If β2 is positive, this signifies that
a “lucky” positive political capital shock for firm i is associated with an increase in the outcome
variable of interest relative to another firm j that experienced an “unlucky” negative political capital
shock during the same election cycle. The null hypothesis for all of our tests is that β2 = 0; that
is, political capital shocks have no effects on our outcome variables.
We also identify a subset of firms as sensitive to economic policy using a procedure that we
describe in detail below and define an indicator variable, Policy Sensitive, to take a value of 1 if
the firm is sensitive and 0 otherwise. We then use a triple-difference framework to study whether
the effects of these political capital shocks differ for firms that are more sensitive or less sensitive
to policy uncertainty. Formally, we estimate the following model:
Outcomei,t = α+ β1PostElectiont + β2Capital Shocki,t × PostElectiont+ (2)
+ β3Posti,t × Policy Sensitivei,t + β4Postt × Policyi,t × Capital Shocki,t+
+ Γ′Controlsi,t + Firm× ElectionCycle FE + εi,t ,
In this specification, the coefficient β4 captures the differential effect of being policy-sensitive
14Julio and Yook (2012) and Durnev (2010) also use a differences-in-differences framework to estimate the effectsof political uncertainty on firm outcomes; however, most of their dependent variables and settings are very differentthan the variables and settings we consider in this paper.
15
on outcomes given the same political capital shock. If, for the sake of argument, both β2 and β4
are negative, than policy sensitive firms had an even larger negative reaction in the outcome to the
same political capital shock than their policy-neutral peers.
3.3 Data
3.3.1 Political Connections Data
Firms contribute money to political candidates in the United States through legal entities known as
Political Action Committees (PACs). PACs solicit contributions from employees of the sponsoring
firm and donate these contributions to one or more political candidates.15 Rather than donating
money directly to candidates’ personal accounts (which is illegal in the United States), firms’ PACs
typically donate money to another PAC set up by a candidate for elected office (known as “Election
PACs”). Firm PAC contributions to Election PACs therefore constitute our measure of political
connectedness linking firms to candidates for elected office.16
We obtain election contribution and election outcome data from the U.S. Federal Election
Commission (FEC) for all federal elections from 1998-2010.17 Panel A of Table 1 presents summary
statistics for our political contributions and election data. United States general elections are held
annually in November. However, all House of Representative and Senate general elections occur in
even numbered years, while Presidential elections occur in years divisible by four.
3.3.2 Options Data
We obtain daily implied volatility data from OptionMetrics from 1997-2011. OptionMetrics com-
putes implied volatility from at-the-money call options using the Black-Scholes model. We use data
for call options with 1 - 6 month maturities. All results are robust to using put options. Panel A
of Table 1 presents summary statistics for our implied volatility data. All of our implied volatility
15Importantly, most decisions regarding which candidates to support are typically left to one or more officers ofthe sponsoring company and frequently to a political specialist such as the PAC chair.
16Firm PAC contributions to Election PACs are legally capped at $10,000 per election cycle.17FEC data is transaction-level data organized by election cycle. Political contribution data is available from the
FEC, the Center for Responsive Politics, or the Sunlight Foundation. The latter two organizations are non-partisan,non-profit organizations who assemble and release government datasets to further the public interest.
16
tests use daily data from six months prior to federal election dates to six months after the election
takes place.
3.3.3 Credit Default Swap Data
We obtain daily CDS data from Markit from 2001 to 2011. Since CDS spreads are not available
prior to 2001, all tests involving CDS spreads only focus on election cycles from 2002 to 2010.
We focus on 1-year, 5-year, and 10-year CDS spreads on senior unsecured U.S.-dollar-denominated
debt. Following Hanouna, Ovtchinnikov, and Prabhat (2014), we take the natural log of the CDS
spread for each firm and use this as a dependent variable in our tests. Panel A of Table 1 presents
summary statistics for our (untransformed) CDS data. All of our CDS tests use daily data from
six months prior to federal election dates to six months after the election takes place.
3.3.4 Economic Policy Uncertainty Data
We use the monthly Economic Policy Uncertainty (EPU) index created by Baker, Bloom, and Davis
(2015) as a proxy for aggregate economic policy uncertainty. This index measures the frequency
of articles that mention economic or policy related uncertainty in 10 major US newspapers. The
authors suggest that this series predicts declines in aggregate investment and output along with
increases in volatility.
3.3.5 Other data
We also obtain quarterly accounting data from COMPUSTAT, daily stock returns from CRSP, and
stock return factors from Ken French’s website. Definitions of all variables used in our tests are
contained in Appendix A.
4 Results
The argument we outlined in section 3 consists of four main components. First, firms differ from
one another in the cross-section and the time series in the extent to which they are sensitive to
17
uncertainty related to economic policy. Second, all else equal, the marginal value of an extra
political connection is larger for policy-sensitive firms, and hence, these firms are more likely to
make campaign contributions relative to policy-neutral firms. Third, shocks to firms’ political
capital bases will lead to changes in firm outcomes following elections. Finally, to the extent that
firms make campaign contributions in support of their privately-optimal policy objectives, policy-
sensitive firms will respond more forcefully to a positive political capital “shock” (i.e. a marginal
political connection) than policy-neutral firms. These arguments are discussed in detail in the
subsections below.
4.1 Firms’ Sensitivities to Economic Policy Uncertainty
Our first task is to document that firms differ in their sensitivities to economic policy uncertainty
(EPU). We seek to understand, among other things, what fraction of firms have a high sensitivity
to EPU, how persistent this sensitivity is across time, and whether sensitive and less-sensitive firms
differ significantly along observable measures such as size, leverage, and industry classification.
In order to examine these questions (and the other questions in this paper), we need to define a
time-varying, firm-specific measure of firms’ sensitivity to economic policy uncertainty.
We use the Economic Policy Uncertainty index developed by Baker, Bloom, and Davis (2015)
as our primary measure of economic policy uncertainty.18 The EPU index is an aggregate time-
series index that measures the frequency of newspaper articles containing words which indicate
uncertainty about economic policy.19 We are not interested in the index level per se, but rather in
agnostically identifying firms that seem to be the most sensitive to economic policy uncertainty.
To identify policy-sensitive firms, we run OLS regressions of the Baker, Bloom, and Davis (2015)
EPU index on each firm’s monthly stock returns in the 18 months leading up to an election. We
18Other measures of economic policy uncertainty exist as well. For example, Whited and Leahy (1996) and Bloom,Bond, and Reenen (2007) examine the link between general uncertainty and investment and use share price volatilityas a firm-specific measure of uncertainty. However, this measure seems to be too general to capture policy-specificuncertainty as opposed to other types of uncertainty. Several authors also use elections to measure time periods whenpolicy uncertainty is high (see, e.g., Gao and Qi (2013) and Julio and Yook (2012)), but this measure cannot be usedto produce ex-ante (i.e. pre-election) cross-sectional variation in policy uncertainty sensitivity at the firm level.
19Brogaard and Detzel (2015) study the general asset-pricing implications of this index and conclude that EPU isan important risk factor for equities. In contrast, we look at the correlation between firms’ stock returns and thisindex at various point in time to identify firms that are sensitive to policy uncertainty.
18
run a separate regression for each firm and each election cycle, so our measure of policy sensitivity
is defined at the firm-election cycle level. We then extract the p-value of the regression coefficient
on the EPU index. We define a firm as being sensitive to economic policy uncertainty during a
given election cycle if the p-value is less than or equal to 0.1. In other words, we define a firm as
being policy-sensitive if the absolute value of its loading on the EPU index is large, regardless of
whether the loading is positive or negative.
Using this procedure has benefits and drawbacks. The main benefit is that we are able to
be agnostic about exactly how uncertainty and returns are related. However, there are at least
two potential drawbacks: (i) we could simply be picking up random variation that is unrelated
to policy sensitivity, and (ii) we are relying on a short time series (18 monthly observations) for
each estimation, which can lead to power problems. However, if we were simply picking up random
variation, we would expect that in a large sample of firms, we would find that roughly 10% of
firms would be sensitive to EPU since this is the p-value cutoff we have defined. Instead, however,
we find that significantly greater than 10% of all firm-cycle observations have p-values lower than
0.1. Hence, we do not appear to be capturing purely noise. Furthermore, all of our main results
remain quantitatively and qualitatively unchanged when we include the market, size, value, and
momentum factors as controls. Finally, all of our main results are robust to varying our sensitivity
cutoff (p-value < 0.1) and the length of our estimation period (18 months).
Panel B, C, and D of Table 1 present summary statistics regarding the fraction and type of firms
that are policy sensitive according to the measure described above. Panel B shows that 18% of the
firm-years in our sample appear to be policy sensitive. Panel B also shows that there is time-series
variation in the fraction of firms that are defined as sensitive to EPU; for example, the 2008 and
2004 political cycles have the largest proportion of sensitive firms/years (48 and 22% respectively)
while 2010 has the lowest proportion (8%).20
Panel C examines the potential persistence of policy sensitivity. In particular, it may be that
20Since our EPU correlations are computed using 18 months worth of data prior to each election, the increasein sensitivity in 2008 is not likely to be simply capturing the U.S. government’s response to the recent financialcrisis since it is computed using data from May 2007 – October 2008, which was roughly the beginning of the U.S.government’s response to the crisis.
19
some firms (or industries) are always policy-sensitive in every election cycle, while other firms are
never policy-sensitive in any election cycle. However, Panel C shows that this does not appear to
be the case; in fact, there are fewer cases of “persistent” policy sensitivity than we would expect
even if policy sensitivity were i.i.d. across each election cycle period. Finally, Panel D presents
the proportion of firms in each of the Fama-French 49 industries. Again, one might think that
some industries may persistently be policy-sensitive, while other industries may almost never be
policy-sensitive. However, Panel D shows that this is not the case: firms in the most policy-sensitive
industry (real estate) are themselves only policy-sensitive approximately 22% of the time, while
firms in the least policy-sensitive industry (agriculture) are still policy-sensitive around 11% of the
time.
We next look at how comparable firms that are policy sensitive are to firms that are not policy
sensitive. Table 2 contains the results of our tests. Panel A examines univariate differences in firm
characteristics such as size, leverage, investment, asset intensity, firm profitability, and Tobin’s Q
(as proxied for by the M/B ratio). The panel shows that policy-sensitive firms tend to be larger,
have higher leverage, and have lower asset intensity (PP&E/assets) relative to policy-neutral firms.
However, while these results are statistically significant, their economic magnitudes are quite small.
For example, policy-sensitive firms have leverage and asset intensity levels that are around 3% and
5% higher and lower than less-sensitive firms, respectively. Hence, while policy-sensitive firms
are not identical to less-sensitive firms along every dimension, neither group stands out as being
substantively different from the other along most observable measures.
We test this proposition more formally in Panel B. This panel presents the results of a logit
regression where the dependent variable is a binary variable taking the value of one if a given firm
is policy-sensitive in a given election cycle, and zero otherwise. Our independent variables are the
same firm characteristics that we studied in Panel A. Panel B shows that with the exception of
book leverage, none of the variables in Panel A appear to be strongly correlated with whether or
not a firm is policy-sensitive in a given election cycle.
The results we have presented thus far indicate that policy uncertainty sensitivity varies both
20
within election cycles and within firms. In particular, the lack of persistence within firms and
the relatively similar observable characteristics of sensitive versus non-sensitive firms suggest that
policy sensitivities most commonly represent distinct “shocks” that are specific to a given election
cycle. While policy sensitivities are not determined randomly, this evidence suggests that it is
unlikely that the effects we document elsewhere are purely driven by “fundamental” differences
between policy-sensitive and policy-neutral firms.
4.2 Policy-Sensitive Firms and Campaign Contributions
Having documented that some firms are more sensitive to policy uncertainty than others, we now
turn to our conjecture that policy-sensitive firms will be more likely to donate (and/or will give
larger amounts) relative to policy-neutral firms. Intuitively, the marginal value of a political con-
nection should be larger for policy-sensitive firms, creating a stronger incentive for policy-sensitive
firms to donate money to candidates for elected office.21
We examine this proposition formally in Table 3, where we regress policy uncertainty sen-
sitivity on a firm’s total political contributions. The main independent variable of interest is
Policy Sensitive, which is a binary variable that takes the value of one if a firm is policy-sensitive
in a given election cycle, and is zero otherwise. Column (1) in the table documents a positive uni-
variate relationship between firms’ policy sensitivity and their total (log) campaign contributions:
on the whole, policy-sensitive firms appear to donate more to political candidates than policy-
neutral firms. In columns (2), (3), and (4), we introduce firm and industry-election cycle fixed
effects and add firm-level control variables such as (log) size, leverage, profit margin, and the firm’s
scaled cash holdings. In all three specifications, the relationship between policy sensitivity and
total political contributions persists: policy-sensitive firms donate more to candidates for elected
office than non-policy sensitive firms in a given election cycle. In columns (5) and (6), we further
split each firm’s political contributions into contributions made to candidates in close elections
and contributions made to candidates in other (non-close) elections. The table shows that while
21Throughout the paper, we are agnostic as to whether contributing firms are attempting to (i) directly influenceelection outcomes, (ii) secure influence with politicians conditional on the politician winning their respective races,or (iii) both.
21
firms contribute more to both types of races when they become policy-sensitive, the increase in
contributions seems to be larger in close election races.
If becoming policy-sensitive leads firms to increase their campaign contributions to candidates
in close elections, one might be concerned that these firms may be able to forecast election outcomes
more accurately than non-policy sensitive firms (which donate less). If firms are able to systemat-
ically predict the winners of close elections, this would invalidate the main identifying assumption
of our remaining empirical analysis since close election outcomes would not be unpredictable at the
time of firms’ donations. We test for this possibility by using our main political shock variable (the
net number of close-election victors supported by each firm) as the dependent variable in the final
two specifications in Table 3. Columns (7) and (8) show that policy-sensitive firms do not appear
to have better forecasting power than their policy-neutral peers when it comes to predicting the
winners of close U.S. congressional elections.
4.3 Implied Volatility
We are now ready to explore the link between political donations, election outcomes, and firms’
subsequent risk-taking.
We begin by examining the implied volatility of firms’ options following political capital shocks.
A number of recent studies have examined the impact of political capital shocks on firms’ stock
returns (see, e.g., Cooper, Gulen, and Ovtchinnikov (2009), Cohen, Diether, and Malloy (2013),
Belo, Gala, and Li (2013), and Addoum, Delikouras, Ke, and Kumar (2014)). Nearly all of these
studies find that positive political capital shocks are associated with higher subsequent firm returns.
To our knowledge, however, no one has linked the results on returns to changes in firms’ risk-
taking. In particular, if we assume that long-run returns are increasing, then standard asset pricing
models such as the CAPM will predict that the increase in expected returns should be caused
by a corresponding increase in the firm’s risk. In light of the existing literature on returns, this
logic suggests that we should expect to see an increase in firms’ risk-taking (as proxied by implied
volatility) if the standard relationship between risk and return continues to hold in our sample.
22
However, Panel A of Table 4 shows that implied volatility decreases for “lucky” firms relative
to “unlucky” firms following election dates. The first six columns in the table examine the implied
volatility on one-month, three-month, and five-month options. Columns (1), (3), and (5) contain
the results from our baseline diff-in-diff setup, while columns (2), (4), and (6) add the underlying
firm’s daily stock return and the stock return on the firm’s value-weighted industry as control
variables.22 These columns show that the drop in idiosyncratic volatility for “lucky” firms is quite
large; for example, the implied volatility on one-month call options declines by approximately
12% following elections for “lucky” firms (Close Win Dummy = 1) relative to “unlucky” firms
(Close Win Dummy = 0). Columns (7) - (9) repeat the analysis using the continuous measure of
a firm’s political capital shock. In unreported results, we confirm that there are opposite effects
of similar magnitude for connections to winning and losing candidates. That is, a firm’s implied
volatility goes down when it gains political connections and increases when it fails to gain such
connections.
Our implied volatility results complement the results of Kelly et al. (2014), who examine the
relationship between implied volatility and political uncertainty. They find that implied volatility
goes up for all stocks during the period just before elections, when political uncertainty is likely to
be the highest. To our knowledge, however, no one has examined the link between implied volatility
and the cross-section of politically-active firms.
We next examine how implied volatility changes differ between firms that are sensitive to pol-
icy uncertainty versus those that are not. Formally, we use a difference-in-difference-in-difference
methodology and examine how this effect varies for firms that have a “lucky draw” and are sensitive
to policy uncertainty compared to those firms that are sensitive to policy uncertainty and have an
“unlucky draw.” Panel B of Table 4 presents this analysis. The primary coefficient of interest in
specifications (1) – (6) is Post×Policy×CloseWinDummy and Post×Policy×NetCloseWins
in specifications (7) – (9), which captures the differential effect for policy-sensitive firms. The triple
interaction term is negative and highly significant in all specifications. The magnitude of the triple-
22Industry returns are computed using three-digit SIC codes. All results in the table are robust to different industrydefinitions such as one-digit or four-digit SIC codes, Fama-French 49 industry definitions, or GICS definitions.
23
difference estimator is larger than the simple difference-in-difference estimator in all specifications,
which suggests that a large fraction of the reduction in implied volatility comes through better con-
nections to politicians in times when the firm is more sensitive to policy uncertainty. For example
in the 5 month option specifications ((5) and (6)), the connection effect for policy-sensitive firms
is 2.5 larger than for non-sensitive firms (-.0572 vs. -.0220).
4.4 Credit Default Swaps
Our second measure of firm riskiness is credit default swaps. CDS spread results are contained in
Table 5. We proceed in the same fashion as for implied volatility — we first conduct a difference-
in-difference analysis for firms that have “lucky draws” and firms that have “unlucky draws” before
and after the election and next perform a triple difference analysis to see how these effects vary
for policy-sensitive firms versus policy-neutral firms. An increase in CDS spreads indicates an
increase in the (expected) credit risk associated with a firm, while a decrease in CDS spreads
indicates a decline in expected credit risk. Panel A presents the simple difference-in-difference
results. Consistent with our previous results, we find that “lucky” shocks to political capital are
associated with lower firm riskiness. The first six columns in the table examine the effects of
political capital shocks on one-year, five-year, and 10-year CDS spreads. Columns (1), (3), and (5)
contain the results from our baseline difference-in-difference specification, while columns (2), (4),
and (6) add a host of control variables to the specification. All six columns shows that CDS spreads
drop significantly for “lucky” firms in the six months following U.S. federal elections. The drop in
firms’ expected credit risk for “lucky” firms is substantial; for example, one-year log CDS spreads
decline by more than 30% for “lucky” firms (Close Win Dummy = 1) relative to “unlucky” firms
(Close Win Dummy = 0).
Panel B presents the results of the triple-difference analysis. Consistent with our analysis of
implied volatility, we find that there is a larger reduction in risk for policy-sensitive firms that had
a “lucky draw” than for policy-neutral firms compared to their peers that had “unlucky draws.”
The economic size of this differential impact is even larger than the effect we found for implied
24
volatility: the smallest relative difference between policy-sensitive and policy-neutral firms is about
3.5 (specification (3)), while the largest difference is 7.5 (specification (2)). These results strongly
suggest that most of the reduction in CDS spreads comes from increases in firms’ political capital
stocks at times when the firm is more exposed to policy uncertainty.
In a contemporaneous paper, Hanouna, Ovtchinnikov, and Prabhat (2014) also examine the
link between political connections and CDS spreads. However, Hanouna et al. (2014)’s tests are
primarily focused on the more general link between political activity and CDS spreads. In essence,
they compare CDS spreads between firms that invest in political capital and firms that do not invest
in political capital, and find that CDS spreads are lower for firms that invest in political capital. In
contrast, we focus on the impact of a political capital shock on risk-taking across a subset of firms
that are all investing in political capital, and we focus on how firms’ post-election risk-taking differs
among policy-sensitive and policy-neutral firms. We also examine the effects of political capital
shocks on the term structure of CDS spreads, which Hanouna et al. (2014) do not examine. In
addition to addressing different questions, the two papers also have different identification strategies
and interpret their results differently. Hence, while both sets of results suggest that CDS spreads
are sensitive to a firm’s investment in political capital, we view our results as being both distinct
and complementary to the results in Hanouna et al. (2014).
4.5 The Term Structure of Political Capital Shocks
While a large literature has found that political capital shocks have an immediate impact on stock
prices, there is no current evidence in the literature regarding the term structure of political capital
shocks. To what extent are the benefits of political capital shocks transitory instead of permanent?
What is the expected half-life of a positive political capital shock? The goal of this section is to
provide some preliminary answers to these questions.
Figures 2 and 3 plot the term structure of political capital shocks for implied volatility and
CDS spreads, respectively. There are two takeaways from the figures. First, the impact of political
capital shocks appears to be long-lasting. For example, Figure 3 shows that 10-year CDS spreads
25
respond (negatively) to positive political capital shocks almost as much as 1-year CDS spreads.
This suggests that the market anticipates the effects of political capital shocks to be long-lasting.
Second, the figures shows a pronounced upward “slope” effect for “lucky” firms. That is, relative to
unlucky firms, lucky firms’ long-dated CDS spreads and option-implied volatilities respond slightly
less emphatically to positive political capital shocks relative to their short-dated brethren.
Teasing out the “half-life” of political capital shocks is not straightforward, but a simple credit
risk example can help to shed light on how the two results in the graph are informative about
the term structure of political capital shocks. Suppose the annual probability of a credit event is
10%. Then the probability of an event occurring in the next five years is 1 − (1 − 0.10)5 ≈ 0.41
(assuming independence across years, which we assume for the purpose of simplicity despite its
unrealistic nature).23 Now suppose that the probability of an event occurring in the first year only
jumps from 0.10 to 0.05. The new probability of an event occurring within five years now drops
to 1 − (0.95 × (1 − 0.1)4 ≈ 0.38. In contrast, suppose that the probability of an event occurring
now decreases to 0.05 not only in the first year, but across all five years. Then the cumulative
probability that an event occurs within five years is 1 − (1 − 0.05)5 ≈ 0.23. In other words, the
difference between the original default probabilities (1-year = 0.1; 5-years = 0.41) and the new
default probabilities (temporary case: 1-year = 0.05, 5-years = 0.38; permanent case: 1-year =
0.05, 5-years = 0.23) is significantly greater when the shock to default probabilities is long-lasting
in nature, particularly at the long end of the maturity spectrum. As such, this implies that a long-
lasting political capital shock should lead to a relative steepening of CDS slopes when compared
with a short-lasting shock.24
This example highlights two points. First, the fact that CDS spreads drop significantly even
for long-dated contracts suggests that the market anticipates a large downward change in the
probability of default extending for periods beyond a single year. In the example above, when
the political capital shock only affects default probabilities in the first year, the effect on 5-year
default probabilities is muted. In contrast, figure 3 shows that there is a very large change in
23In future iterations, we will use a hazard model for this example.24Importantly, in absolute terms, the CDS slope is now flatter; in the original case, the slope was 0.41 - 0.1 = 0.31;
in the case of a permanent shock, the absolute slope is now 0.23 - 0.05 = 0.18.
26
CDS spreads for long-dated contracts, suggesting that the effects we are capturing have lengthy
maturities. Second, in the example above, a sustained change in default probabilities creates a
steeper change in CDS slopes than a single-year change. Figure 3 shows that CDS slopes indeed
steepen modestly for “lucky” firms following positive political capital shocks. As such, figure 3
presents visual evidence that the effects of a political capital shock appear to have long-dated
effects on expected firm riskiness.
We test this proposition formally in Table 6. Panel A of the table presents results on the effect
of political capital shocks on implied volatility slopes. Specifications (1) – (3) repeat the difference-
in-difference estimation, while specifications (4) – (6) repeat the triple difference specification.
Specifications (1) – (3) suggest that there is in fact an upwards sloping term structure of these
shocks unconditionally. However, specifications (4) – (6) suggest that these patterns do not persist
for the firms that are policy sensitive. Panel B of the table presents results on the effect of political
capital shocks on CDS slopes. In contrast with the implied volatility slopes, in all six specifications
in the panel, “lucky” firms’ CDS slopes steepen by as much as 15% relative to “unlucky” firms
following close elections. These effects are stronger for policy-sensitive firms, and may in fact
completely explain these results, since the coefficients on Post×CloseWinDummy lose economic
and statistical significance in the triple-difference estimation. When combined with the results from
Table 5, these results support the idea that the effects of political capital shocks have a long-term
(expected) effect on firms’ credit risk.
4.6 Firms’ Investment and Operating Decisions
Tables 4 and 5 suggest that market-driven proxies for firm risk-taking decline following positive
political capital shocks, and decline particularly strongly (in magnitude) in the case of policy-
sensitive firms. However, these tables do not shed light on how firms’ risk-taking might be changing
following positive political capital shocks. To address this question, Tables 7 and 8 examine how
firms’ investment, leverage, R&D spending, profitability, and operational performance respond to
“lucky” political capital shocks. As in tables 4 and 5, we begin by examining the differential
27
response of “lucky” winners versus “unlucky” losers following the outcomes of close elections. We
then further split our sample based on whether firms are particularly sensitive to economic policy
uncertainty during a given election cycle.
Table 7 examines how firms’ investment, leverage, and R&D spending behavior respond to po-
litical capital shocks. The results in Panel A suggest that firms do not appear to significantly adjust
their investment, leverage, or R&D spending policies in response to a political capital shock: the
interaction term between the post-election and “lucky” close-election dummy variables is statisti-
cally zero in every specification. The fact that we do not find a differential post-election change in
leverage between “lucky” and “unlucky” firms contrasts with Claessens, Feijen, and Laeven (2008),
who find that leverage increases following positive political capital shocks in Brazil. Our findings
of no differential effects on investment and R&D also contrast with Do, Lee, and Nguyen (2013),
Ovtchinnikov, Reza, and Wu (2014), and Kim (2015), who report evidence that political capital
shocks have significant effects on investment and innovation (though in different directions).
However, when we segment our sample further based on firms’ differing sensitivity to economic
policy uncertainty, we find that policy-sensitive firms’ investment and leverage do respond strongly
to political capital shocks. In particular, Panel B shows that policy-sensitive firms respond to
a “lucky” political capital shock by increasing investment and decreasing leverage following close
election outcomes, though we do not find any effects on firms’ R&D spending. Our investment
and leverage results are economically large; holding all else equal, “lucky” policy-sensitive firms’
investment increases by about 11% and leverage decreases by about 2% relative to similarly “lucky”
firms that are less sensitive to economic policy shocks. These results are even stronger when
we compare policy-sensitive “lucky” winners versus other policy-sensitive firms that experienced
“unlucky” political capital shocks. Hence, the resolution of political uncertainty appears to have a
significantly different impact on the investment and leverage choices of policy-sensitive firms based
on whether these firms’ experienced positive or negative political capital shocks.
Table 8 extends the tests in Table 7 to examine how firms’ operating performance and profitabil-
ity respond to political capital shocks. Panel A present difference-in-difference results, while Panel
28
B presents triple-difference results. The results in Panel A suggest that “lucky” firms experience
higher sales, higher returns on assets, and higher M/B ratios than “unlucky” firms following close
election outcomes. These results are largely in line with the existing literature.25
However, Panel B shows that these results are largely driven by policy sensitive-firms. For
example, Columns (1) and (3) of Panel B show that the increases in sales and M/B documented in
Panel A accrue almost exclusively to policy-sensitive firms who experience a “lucky” political cap-
ital shock. Furthermore, Panel B shows that “lucky” policy-sensitive firms also experience higher
asset growth, higher gross profit margins, and higher net profit margins than other “lucky” winners.
These effects are also economically significant; for example, the COGS/Sales ratio decreases by 6%
for “lucky” policy-sensitive firms relative to other lucky firms, and by 10% relative to “unlucky”
policy-sensitive firms. Collectively, the results in Table 8 indicate that shocks to political connect-
edness are associated with significantly stronger future operating performance and profitability,
particularly for policy-sensitive firms. While other studies have shown that sales positively respond
to an increase in political connectedness, all of the other results in Table 8 are new.
5 Alternative Mechanisms
Firms’ differential sensitivity to economic policy uncertainty appears to be a driving factor behind
many of our results. However, the literature has proposed alternative theories that link political
capital shocks to firm risk-taking. One possibility is that firms establish political connections to
insure themselves against future shocks — a bailout story. One would expect that if this were
the case, then an increase in political connectedness would be followed by an increase in firms’
ex-ante risk-taking. A second possibility is that firms establish political connections to increase
the probability of winning future government contracts (e.g. Cohen and Malloy (2014)). Since
the government may be less likely to award contracts to a distressed firm, these “government-
dependent” firms may want to reduce risk-taking following a positive shock to political capital in
order to lock in the expected gains from future government contracts. We test these both of these
25For example, Akey (2015), Goldman, Rocholl, and So (2013), Tahoun (2014), and Amore and Bennedsen (2013)find evidence that sales growth increases following an increase in political connectedness.
29
potential mechanisms below.
5.1 Bailout Likelihood
One would expect that if firms were establishing political connections to hedge against adverse
future states that they would take on additional risk when they were able to form more political
connections. Acharya, Pedersen, Philippon, and Richardson (2010) propose a measure of ex-ante
risk taking (systemic risk) in the context of the financial crisis — Marginal Expected Shortfall
(MES). MES measures a firm’s expected loss given that the market has a large negative return. A
smaller (more negative) value of MES indicates higher ex-ante risk. To examine the link between
political capital shocks and MES, we run the same differences-in-differences specification as in
previous tests. Table 9 presents the results of this analysis. All specifications indicate that MES
increases for better politically connected firms relative to the less connected peers in post election
years. This result in and of itself cannot rule out a “bailout” story: risk-taking could still be
increasing while the “price” of risk could be decreasing more, leading to an overall decrease in
MES despite an increase in risk-taking. However, we find little evidence of increased risk-taking
elsewhere in our tests; for example, our differences-in-differences tests show little evidence that
firms increase investment or leverage following positive political capital shocks. Hence, while we
cannot rule out a “bailout” story, we view this story as being unlikely to explain our results.
5.2 Government Contractors
The other potential channel that has been described in the literature is a “sticky-government-
contracting” channel. Intuitively, firms that rely more on government contracts for their sales
have a stronger interest in supporting political candidates who can help them to earn future sales.
Conditional on having established political connections, these firms also have a strong incentive to
stay in business – loosely speaking, if a firm donates money to candidate X and candidate X is then
elected, the firm will not want to do anything to jeopardize its ability to profit from its relationship
with candidate X going forward. As such, this story implies that large government contractors
may respond to a positive political capital shock by reducing their risk-taking in anticipation of
30
obtaining future government contracts. Alternatively, firms experiencing a positive political capital
shock (and hence, a higher probability of obtaining government contracts) may simply kick back
and enjoy the “quiet life,” since their future earnings streams are less subject to market competition.
Cohen and Malloy (2014) find evidence consistent with this argument: they find that investment
is lower and operating performance is worse at government-dependent firms.
To test this story, we download segment data from COMPUSTAT and classify firms as being
“government-dependent” in a given quarter if they list the U.S. Government as one of their operating
segments. We then examine whether our previous results on risk-taking are being driven primarily
by government-dependent firms.
Before continuing, it is important to note that we identify government-dependent firms dif-
ferently than Cohen and Malloy (2014). They examine regulatory filings to find firms who ob-
tain more than 10% of sales from the U.S. government, whereas we simply examine firms that
have a separate operating segment for government sales. While it is not clear which approach is
more “correct,” it is possible that our identification procedure mis-classifies government-dependent
firms as non-government-dependent firms. This could bias our results against finding an effect for
government-dependent firms, and may also reduce our power to detect an effect for such firms. As
such, we view our findings below as simply descriptive.
Table 10 contains the results of our tests. In the table, we have chosen to focus on CDS
spreads; however, we obtain similar results for our other tests. Columns (1) - (3) of the table
show that the CDS spreads of government-dependent firms do indeed respond more negatively to
positive political capital shocks than non-government-dependent firms (e.g. the loading on the
triple-interaction term Post× Close Election Dummy ×Gov Contractor is negative). However,
this result is not statistically significant. Furthermore, the table shows that our primary effect
does not appear to be coming from government-dependent firms: columns (7)-(9) confirm that our
main result goes through even after excluding government-dependent firms from our sample. In
fact, columns (4)-(6) show that even within the sample of government-dependent firms, we obtain
qualitatively similar results to our prior results using all firms; that is, we still find that CDS
31
spreads drop more following elections for firms experiencing a positive political capital shock, with
similar magnitudes to our findings for non-government-dependent firms. Hence, our results do not
appear to be fully explained by the “sticky-government-contracting” hypothesis.
5.3 Firm versus Industry Effects
We also run a series of tests to ensure that we are not simply picking up industry-specific effects in
our results. For example, it may be that firms in the defense industry tend to support Republicans
while firms in the technology industry tend to support Democrats. One concern might be that
we are simply capturing the fact that, say, defense contractors tend to vote for Republican party
candidates and Republicans happened to fare well in a given election cycle. In this case, all defense
contractors may have large “political capital shocks” in years where Republican candidates fare
well, whereas the same firms will have low political capital shocks in years where Republicans do
poorly. While this scenario does not violate our primary identifying assumptions, it does present an
alternative explanation for our results on firm risk-taking. In particular, this scenario suggests that
our “political capital shocks” may not be stemming from actual shocks, but may instead simply be
picking up longstanding political affiliations between certain industries and certain political parties.
Table 11 tests this idea formally by examining whether our main results in Tables 4 through 8
are robust to different specifications that examine the effects of political capital shocks on firm
risk-taking by looking within industries as opposed to within firms. In particular, we replace the
firm-election cycle fixed effects that were used in Tables 4 through 8 with a combination of firm fixed
effects and industry-election cycle fixed effects (where industry definitions are based on Fama and
French (1997)’s 49-industry classification system). These tests examine whether positive political
capital shocks are associated with lower firm risk-taking within a given industry and election cycle.
In other words, even if all defense contractors had positive political capital shocks in a given cycle
(say, 2002), this test tells us whether defense contractors that received more positive political
capital shocks than other defense contractors in this election cycle are still more likely to reduce
their risk-taking following the outcome of the election.
32
Table 11 shows that the answer to this question is yes: all of our main results are qualitatively
and quantitatively similar regardless of whether we look within industry-election cycles (as in
Table 11) or within firm-election cycles (as in our earlier tests). Hence, the effects we document in
this paper are not simply picking up firms from certain industries or longstanding industry-specific
political affiliations.
6 Robustness
We perform a variety of tests to ensure the robustness of our main results. We begin by constructing
placebo tests for our main difference in difference specifications where we artificially move the actual
election dates in our sample forward to verify that there are no “pre-trends” in the data for our two
groups of firms. These robustness checks are detailed in our Internet Appendix and are available
to review upon request.
We also verify that the parallel trends assumptions behind our differences-in-differences spec-
ifications are met empirically. One major concern, in particular given that much of our data is
observed on a daily frequency, is that news about the likelihood of close election outcomes becomes
available and some firms strategically re-allocate their contributions to support candidates that are
now more likely to win. If some firms do this, but other firms do not, this could create a violation
of the parallel trends assumptions. It is not immediately obvious that this is a problem because
firms frequently make small contributions to the same candidate at various time periods during
an election cycle, when the outcome is less likely to be predictable. However, in order to address
this concern, we recompute our connection measures excluding all contributions made in the two
months leading up to the election. On the one hand, this filter causes us to exclude about 20% of
the contributions that firms make, but on the other hand, the correlation between our connection
measures with and without these contributions is 80-90%. We replicate our main findings using
these results and find that nearly all of our results are quantitatively unchanged. Tables 1 and 2
of the Internet Appendix reproduce our results.
Finally, one potential concern for our MES results is that the financial crisis is driving our
33
results. We address this concern by removing the 2008 political cycle (years 2008 and 2009) from
our sample and replicating our MES results. All of the results in Table 9 are comparable in
significance and magnitude excluding the crisis. Table 3 of the Internet Appendix presents this
analysis.
7 Conclusion
We link the cross-section of firms’ sensitivities to economic policy uncertainty to their subsequent
political activity and post-election risk-taking behavior. We first show that firms with a high sensi-
tivity to economic policy uncertainty donate more to candidates for elected office than less-sensitive
firms. Using a sample of close U.S. congressional elections, we then show that plausibly exogenous
positive shocks to policy-sensitive firms’ political capital bases produce large subsequent changes in
these firms’ investment, leverage, operating performance, CDS spreads, and option-implied volatil-
ity. These effects are weaker in magnitude among less policy-sensitive firms, suggesting that the
marginal value of a political connection is increasing in firms’ ex-ante sensitivity to policy uncer-
tainty. We also examine the term structure of credit risk and implied volatility following political
capital shocks and find that these shocks are expected to be long-lasting in nature. Our results
cannot be explained by moral hazard or government contracting arguments and represent the first
attempt in the literature to shed light on the relationship between firms’ policy sensitivities and
their subsequent risk-taking behavior. While there are no existing theories (to our knowledge) that
link policy uncertainty to political capital and political capital to risk-taking, we view this as an
interesting prospective area for future research.
34
References
Acemoglu, Daron, Simon Johnson, Amir Kermani, James Kwak, and Todd Mitton (2013), “TheValue of Connections in Turbuelent Times: Evidence from the United States.” NBER workingpaper.
Acharya, Viral, Lasse Heje Pedersen, Thomas Philippon, and Matthew P. Richardson (2010), “Mea-suring Systemic Risk.” Working Paper, New York University.
Addoum, Jawad, Stefanos Delikouras, Da Ke, and Alok Kumar (2014), “Under-Reaction to PoliticalInformation and Price Momentum.” Working paper, University of Miami.
Agarwal, Rajesh, Felix Meshke, and Tracy Wang (2012), “Corporate Political Contribution: In-vestment or Agency?” Business and Politics, 14.
Akey, Pat (2015), “Valuing Changes in Political Networks: Evidence from Campaign Contributionsto Candidates in Close Congressional Elections.” Review of Financial Studies, Forthcoming.
Amore, Mario and Morten Bennedsen (2013), “Political District Size and Family-Related FirmPerformance.” Journal of Financial Economics, 110, 387–402.
Baker, Scott R., Nicholas Bloom, and Steven J. Davis (2015), “Measuring Economic Policy Uncer-tainty.” Working paper, Stanford.
Belo, Francisco, Vito Gala, and Jun Li (2013), “Government Spending, Political Cycles, and theCross Section of Stock Returns.” Journal of Financial Economics, 107, 305 – 324.
Bloom, Nick, Stephen Bond, and John Van Reenen (2007), “Uncertainty and Investment Dynam-ics.” Review of Economic Studies, 74, 391–415.
Borisov, Alexander, Eitan Goldman, and Nandini Gupta (2015), “The Corporate Value of (Cor-rupt) Lobbying.” Reveiw of Financial Studies, Forthcoming.
Boutchkova, Maria, Hitesh Doshi, Art Durnev, and Alexander Molchanov (2012), “PrecariousPolitics and Return Volatility.” Review of Financial Studies, 25, 1111–1154.
Brogaard, Jonathan and Andrew Detzel (2015), “The Asset-Pricing Implications of GovermentEconomic Policy Uncertainty.” Managment Science, 61, 3–38.
Claessens, Stijn, Erik Feijen, and Luc Laeven (2008), “Political Connections and Preferential Accessto Finance: The Role of Campaign Contributions.” Journal of Financial Economics, 88, 554–580.
Coates IV, John C. (2012), “Corporate Politics, Governaance and Value before and after Citizen’sUnited.” Journal of Empirical Legal Studeies, 9, 657–696.
Cohen, Lauren, Karl Diether, and Christopher Malloy (2013), “Legislating Stock Prices.” Journalof Financial Economics, 110, 574–595.
Cohen, Lauren and Christopher Malloy (2014), “Mini West Virginias: Corporations as GovernmentDependents.” Working Paper, Harvard University.
Cook, Thomas D. and Donald T. Campbell (1979), Quasi-Experimentation: Design & AnalysisIssues for Field Settings. Houghton Mifflin, Boston.
35
Cooper, Michael, Huseyn Gulen, and Alexei Ovtchinnikov (2009), “Corporate Political Contribu-tions and Stock Returns.” Journal of Finance, 65, 687–724.
Do, Quoc-Ahn, Yen Teik Lee, and Bang Dang Nguyen (2013), “Political Connections and FirmValue: Evidence from the Regression Discontinuity Design of Close Gubernatorial Elections.”Unpublished working paper.
Do, Quoc-Ahn, Yen Teik Lee, Bang Dang Nguyen, and Kieu-Trang Nguyen (2012), “Out of Sight,Out of Mind: The Value of Political Connections in Social Networks.” Working paper, Universityof Cambridge.
Duchin, Ran and Denis Sosyura (2012), “The Politics of Government Investment.” Journal ofFinancial Economics, 106, 24–48.
Durnev, Art (2010), “The Real Effects of Political Uncertainty: Elections and Investment Sensitiv-ity to Stock Prices.” Unpublished working paper: University of Iowa.
Faccio, Mara (2004), “Politically Connected Firms.” American Economic Review, 96, 369–386.
Faccio, Mara, Ron Masulis, and John McConnell (2006), “Political Connections and CorporateBailouts.” Journal of Finance, 61, 2595–2635.
Faccio, Mara and David Parsley (2009), “Sudden Deaths: Taking Stock of Geographic Ties.”Journal of Financial and Quantitative Analysis, 33, 683–718.
Fama, Eugene F. and Kenneth R. French (1997), “Industry Costs of Equity.” Journal of FinancialEconomics, 43, 153–193.
Fisman, Raymond (2001), “Estimating the Value of Political Connections.” American EconomicReview, 91, 1095–1102.
Gao, Pengjie and Yaxuan Qi (2013), “Political Uncertainty and Public Financing Costs: Evidencefrom U.S. Municipal Bond Markets.” Working paper, Notre Dame.
Goldman, Eitan, Jorg Rocholl, and Jongil So (2009), “Do Politically Connected Boards Add FirmValue?” Review of Financial Studies, 17, 2331–2360.
Goldman, Eitan, Jorg Rocholl, and Jongil So (2013), “Politically Connected Boards and the Allo-cation of Procurement Contracts.” Review of Finance, 22, 1617–1648.
Gulen, Huseyin and Mihai Ion (2015), “Policy Uncertainty and Corporate Investment.” Review ofFinancial Studies, Forthcoming.
Hanouna, Paul, Alexei V. Ovtchinnikov, and Saumya Prabhat (2014), “Political Investments andthe Price of Credit Risk: Evidence from Credit Default Swaps.” Working Paper, The Universityof Auckland.
Holthausen, Duncan M. (1979), “Hedging and the Competitive Firm Under Price Uncertainty.”American Economic Review, 69, 989–995.
Johnson, Simon and Todd Mitton (2003), “Cronyism and Capital Controls: Evidence fromMalaysia.” Journal of Finanical Economics, 67, 351–382.
36
Julio, Brandon and Yongsuk Yook (2012), “Political Uncertainty and Corporate Investment Cy-cles.” Journal of Finance, 67, 45–84.
Kelly, Bryan, Lubos Pastor, and Pietro Veronesi (2014), “The Price of Political Uncertainty: Theoryand Evidence from the Option Market.” Unpublished working paper: University of Chicago.
Khwaja, Asim Ijaz and Atif Mian (2005), “Do Lenders Favor Politically Connected Firms? RentProvision in an Emerging Financial Market.” Quarterly Journal of Economics, 120, 1371–1411.
Kim, Tae (2015), “Does a Firm’s Political Capital Affect its Investment and Innovation?” Workingpaper, Notre Dame.
Murphy, Kevin M., Andrei Shleifer, and Robert W. Vishny (1993), “Why is Rent-Seeking so Costlyto Growth?” American Economic Review Papers and Proceedings, 83, 409–414.
Ovtchinnikov, Alexei, Syed Walid Reza, and Yanhui Wu (2014), “Political Activism and FirmInnovation.” Working paper, HEC Paris.
Pastor, Lubos and Pietro Veronesi (2012), “Uncertainty about Government Policy and StockPrices.” Journal of Finance, 67, 1219–1264.
Pastor, Lubos and Pietro Veronesi (2013), “Political Uncertainty and Risk Premia.” Journal ofFinancial Economics, 110, 201–545.
Rajan, Raghuram G. and Luigi Zingales (1998), “Which Capitalism? Lessons from the East AsianCrisis.” Journal of Applied Corporate Finance, 11, 40–48.
Shleifer, Andrei and Robert W. Vishny (1994), “Politicians and Firms.” Quarterly Journal ofEconomics, 109, 995–1025.
Tahoun, Ahmed (2014), “The Role of Stock Ownership by us Members of Congress on the Marketfor Political Favors.” Journal of Financial Economics, 111, 86–110.
Whited, Toni M. and John V. Leahy (1996), “The Effect of Uncertainty on Investment: SomeStylized Facts.” Journal of Money, Credit, and Banking, 28, 64–83.
37
Appendix A - Variable Definitions
Variable Definition Source
Post Election A binary variable that takes the value of 1 in all timeperiods following an election.
CRSP
Close Wins The number of winning candidates involved in a closegeneral election that a firm donated to prior to the elec-tion
Federal Election Commissionand Authors’ Computation
Close Losses The number of losing candidates involved in a close gen-eral election that a firm donated to prior to the election
Federal Election Commissionand Authors’ Computation
Net Close Wins Close Wins - Close Losses Federal Election Commissionand Authors’ Computation
Post × Net CloseWins
Net Close Wins multiplied by Post Election Federal Election Commissionand Authors’ Computation
Post × Close Wins Close Wins multiplied by Post Election Federal Election Commissionand Authors’ Computation
Post × Close Losses Close Losses multiplied by Post Election Federal Election Commissionand Authors’ Computation
Close Win Dummy A binary variable that takes the value of 1 if Net CloseWins > 0
Federal Election Commissionand Authors’ Computation
Post × Close WinDummy
Close Win Dummy multiplied by Post Election Federal Election Commissionand Authors’ Computation
CAPM Beta A firm’s CAPM Beta CRSP and Authors’ Computa-tion
CAPM Vol The standard deviation of residuals of a from annual Betaregressions
CRSP and Authors’ Computa-tion
MES A firm’s Marginal Expected Shortfall computed followingAcharya et al. (2010)
CRSP and Authors’ Computa-tion
Log1y The natural log of a firm’s 1-year CDS spread MarkitLog5y The natural log of a firm’s 5-year CDS spread MarkitLog10y The natural log of a firm’s 10-year CDS spread MarkitLog5y1y Log5y - Log1y MarkitLog10y5y Log10y - Log5y MarkitLog10y1y Log10y - Log1y MarkitSlope5y1y A firm’s unadjusted 5-year CDS spread minus the same
firm’s 1-year CDS spreadMarkit
Slope10y5y A firm’s 10-year CDS spread minus the same firm’s 5-yearCDS spread
Markit
Slope10y1y A firm’s 10-year CDS spread minus the same firm’s 1-yearCDS spread
Markit
1-month impliedvolatility
The implied volatility of a firm’s 1-month at-the-moneycall options
OptionMetrics
3-month impliedvolatility
The implied volatility of a firm’s 3-month at-the-moneycall options
OptionMetrics
5-month impliedvolatility
The implied volatility of a firm’s 5-month at-the-moneycall options
OptionMetrics
3mo-1mo volatilityslope
3-month implied vol minus 1-month implied vol OptionMetrics
5mo-3mo volatilityslope
5-month implied vol minus 3-month implied vol OptionMetrics
5mo-1mo volatilityslope
5-month implied vol minus 1-month implied vol OptionMetrics
Firm return A firm’s daily (unless otherwise noted) stock return (vari-able: ret)
CRSP
M/B A firm’s market capitalization divided by its lagged bookvalue of equity (variables: prc, shrout, ceqq)
CRSP, Compustat
38
VW ind return The value-weighted return on a firm’s 3-digit SIC indus-try (weights are based on market capitalization)
CRSP
Ln(Size) The natural log of the firm’s book value of assets (vari-able: atq)
Compustat
Investment Quarterly capital expenditures divided by lagged netPP&E (variables: capxy (adjusted), ppentq)
Compustat
R&D spending Quarterly R&D expenditure divided by book assets (vari-ables: xrdq, atq)
Compustat
Book leverage Quarterly book value of debt divided by book assets(variables: dlcq, dlttq, atq)
Compustat
Sales growth Quarter-over-quarter change in a firm’s sales (variable:revtq)
Compustat
Asset growth Quarter-over-quarter change in a firm’s assets (variable:atq)
Compustat
EBITDA growth Quarter-over-quarter change in a firm’s EBITDA (vari-able: oibdpq)
Compustat
Profit margin EBIT divided by sales (variables: oiadpq, revtq) CompustatCOGS Cost of goods sold divided by sales (variables: cogsq,
revtq)Compustat
SG&A Selling, general, & administrative expenses divided bysales (variables: xsgaq, revtq)
Compustat
Cash Cash & cash equivalents divided by book assets (vari-ables: cheq, atq)
Compustat
Current ratio Current assets divided by current liabilities (variables:actq, lctq)
Compustat
Scaled FCF EBITDA divided by lagged net PP&E (variables: oib-dpq, ppentq)
Compustat
Profitability EBITDA divided by assets (variables: oiadpq, atq) Compustat
39
Figure 1: Net Close Wins Histogram
The figure below shows the distribution of NetCloseWins measured from 1998-2010.
40
Figure 2: Implied Volatility Term Structure
The figure below plots the Implied Volatility Term Structure for firms that have an above median draw. The solidline plot the point estimate for implied volatility at different call option maturities. The dashed lines show the upperand lower bounds of a 95% confidence interval.
-0.06
-0.055
-0.05
-0.045
-0.04
-0.035
-0.03
-0.025
1 2 3 4 5 6
Maturity of Call Option (Month)
Call Option Term Structure
Point Estimate 95% CI 95% CI
41
Figure 3: Credit Default Swap Spread Term Structure
The figure below plots the Credit Default Swap Spread Term Structure for firms that have an above median draw.The solid line plot the point estimate for the CDS spread at different contract maturities. The dashed lines show theupper and lower bounds of a 95% confidence interval.
-0.0075
-0.0065
-0.0055
-0.0045
-0.0035
-0.0025
-0.0015
1 5 10
Maturity of CDS Contract (Years)
CDS Spread Term Structure
Point Estimate CI 95% CI 95%
42
Table 1: Summary StatisticsPanel A presents summary statistics for (i) political connections data (taken from Federal Election Commissionfilings), (ii) implied volatility data (OptionMetrics), and (iii) CDS spreads (Markit). Definitions of all variablesdescribed in the panel can be found in the text and Appendix A. Panels B, C, and D report summary statistics forfirms that are sensitive to the Economic Policy Uncertainty (EPU) index of Baker, Bloom, and Davis (2015) basedon our estimation procedure (details of which can be found in the text). Panel B reports the number and proportionof firms in each election cycle that are sensitive to the EPU index as well as the fraction of sensitive firms whose EPUsensitivities are positive and negative, respectively. Panel C reports summary statistics on the number of electioncycles that a given firm is policy-sensitive according to our estimation procedure. Panel D reports summary statisticsregarding the industry distribution of policy-sensitive firms across our sample period (1998-2010).
Panel A — Political Connections, Implied Volatility, and CDS Spreads
Data Type Variable Mean Median Std. Dev. Number
Political Connections Net Close Wins 0.13 0 2.37 7,838Close Wins 2.85 2 3.22 7,838Close Losses 2.71 2 2.77 7,838Total Contributions $118,762 39,950 231,214 5,433Close Election Contributions $16,328 7,000 25,896 3,988Other Contributions $107,469 35,775 210,956 5,398
Implied Volatility 1 month implied volatility 0.4038 0.3495 0.2290 842,1903 month implied volatility 0.3934 0.3453 0.2123 840,0895 month implied volatility 0.3869 0.3423 0.2025 830,8313mo/1mo slope -0.0100 -0.0033 0.0465 840,0895mo/3mo slope -0.0060 -0.0016 0.0246 830,8315mo/1mo slope -0.0159 -0.0058 0.0572 830,831
CDS Spreads 1 year spread 0.0183 0.0035 0.0878 354,6975 year spread 0.0214 0.0078 0.0569 387,28710 year spread 0.0216 0.0094 0.0498 358,3445yr/1yr slope 0.0036 0.0028 0.0255 353,2715yr/1yr slope 0.0006 0.0014 0.0121 356,65810yr/1yr slope 0.0043 0.0043 0.0341 342,762
Panel B — Firm Sensitivity to Economic Policy Uncertainty
Election All Policy-Sensitive Fraction Positive NegativeCycle Firms Firms Sensitive Std. Dev. Sensitivity Sensitivity
1998 10,211 1,463 0.143 0.350 29% 71%2000 9.698 1,248 0.129 0.335 41% 59%2002 8,195 938 0.114 0.318 43% 57%2004 7,376 1,586 0.215 0.411 93% 7%2006 7,462 905 0.122 0.326 32% 68%2008 7,646 3,689 0.482 0.500 7% 93%2010 7,203 568 0.079 0.270 39% 61%
Total 57,791 10,397 0.180 0.382 35% 65%
Panel C — Number of Policy-Sensitive Election Cycles Per Firm(Sample restricted to firms present in all seven election cycles)
Number of Cycles where Firm Empirical Binomial Dist.Firm is Policy-Sensitive Count Distribution (p = 0.180)
0 cycles 840 26.7% 25.0%1 cycle 1,317 41.8% 38.3%2 cycles 761 24.2% 25.2%3 cycles 195 6.2% 9.2%4 cycles 32 1.0% 2.0%5 cycles 4 0.1% 0.3%6 cycles 0 0.0% 0.0%7 cycles 0 0.0% 0.0%
Test: Actual = Binomial Chi-Square p-Value N64.60 < 0.0001 3,14943
Table 1: Summary Statistics (Continued)
Panel D — Number of Policy-Sensitive Firms Per Fama-French 49 Industry
IndustryIndustry Number Count Mean Std. Dev.
Real Estate 47 304 0.224 0.417Computers 35 773 0.210 0.407Electronic Equipment 37 2095 0.208 0.406Non-Metallic and Industrial Metal Mining 28 208 0.202 0.402Measuring and Control Equipment 38 657 0.199 0.4Communication 32 1379 0.198 0.399Precious Metals 27 350 0.191 0.394Machinery 21 1006 0.191 0.393Shipbuilding and Railroad Equipment 25 69 0.188 0.394Chemicals 14 506 0.186 0.389Fabricated Products 20 88 0.182 0.388Candy and Soda 3 150 0.180 0.385Business Services 34 2365 0.179 0.383Transportation 41 794 0.179 0.383Electrical Equipment 22 834 0.179 0.383Trading 48 8344 0.175 0.38Defense 26 63 0.175 0.383Textiles 16 132 0.174 0.381Construction 18 401 0.172 0.378Apparel 10 349 0.172 0.378Restaurants, Hotels, and Motels 44 685 0.171 0.377Insurance 46 1176 0.169 0.375Aircraft 24 149 0.168 0.375Computer Software 36 2549 0.167 0.373Tobacco Products 5 60 0.167 0.376Steel Works 19 458 0.166 0.372Medical Equipment 12 1049 0.166 0.372Recreation 6 271 0.162 0.369Petroleum and Natural Gas 30 1411 0.162 0.368Almost Nothing 49 268 0.160 0.368Printing and Publishing 8 351 0.160 0.367Entertainment 7 444 0.158 0.365Automobiles and Trucks 23 457 0.155 0.363Business Supplies 39 368 0.155 0.362Construction Materials 17 475 0.154 0.361Wholesale 42 1407 0.154 0.361Utilities 31 1043 0.153 0.361Consumer Goods 9 339 0.153 0.361None None 850 0.152 0.359Pharmaceutical Products 13 1965 0.149 0.356Healthcare 11 626 0.141 0.348Shipping Containers 40 100 0.140 0.349Banking 45 4130 0.140 0.347Coal 29 80 0.138 0.347Retail 43 1610 0.137 0.344Beer and Liquor 4 156 0.135 0.342Personal Services 33 404 0.134 0.341Food Products 2 483 0.124 0.33Rubber and Plastic Products 15 205 0.117 0.322Agriculture 1 104 0.106 0.309
44
Table 2: Economic Policy Uncertainty Sensitivity and Firm CharacteristicsThis table presents summary statistics for firms that are sensitive to Economic Policy Uncertainty and for those thatare not sensitive. Panel A presents univariate differences, while Panel B presents results from a Logit analysis with anindicator variable that takes the value of one if a firm has been classified as sensitive to economic policy uncertaintyand 0 otherwise as the dependent variable. Details of this estimation procedure are found in the text. Standarderrors are presented in parentheses. The symbols *, **, and *** denote statistical significance at the 10%, 5%, and1% levels, respectively.
Panel A — Univariate Differences
Book Investment / Market / Net PPE / Profit Return onLn(Size) Leverage Capital Book Assets Margin Assets
Other firms 9.057 0.684 0.052 3.005 0.312 0.104 0.021(0.012) (0.001) (0.001) (0.024) (0.002) (0.005) (0.001)22,668 22,415 19,387 21,480 21,779 20,501 22,115
Policy-sensitive firms 9.282 0.706 0.051 2.973 0.297 0.118 0.021(0.027) (0.004) (0.001) (0.059) (0.004) (0.004) (0.001)4,866 4,796 4,310 4,564 4,657 4,480 4,743
Difference 0.225*** 0.021*** -0.001 -0.031 -0.015*** 0.014 0.000(0.028) (0.004) (0.001) (0.059) (0.004) (0.011) (0.001)
Panel B — Logit Analysis
(1) (2) (3) (4) (5) (6)Policy- Policy- Policy- Policy- Policy- Policy-
Variable Sensitive Sensitive Sensitive Sensitive Sensitive Sensitive
ln(Size) 0.0595** 0.0580 0.00672 0.00356 0.0284 0.0284(0.0282) (0.0467) (0.0373) (0.0437) (0.0484) (0.0586)
Book Leverage 0.610** 0.634 1.003*** 1.008** 1.112** 1.112*(0.276) (0.439) (0.372) (0.453) (0.505) (0.603)
It/Kt−1 -0.546 -0.511 -1.224 -1.203 -1.195 -1.195(0.844) (1.164) (1.084) (1.283) (1.242) (1.421)
M/B -0.00658 -0.00906 0.0115 0.00880 0.0170 0.0170(0.0114) (0.0170) (0.0133) (0.0162) (0.0178) (0.0203)
Profit Margin 0.0187 0.0149 0.0529 0.0488 0.0781 0.0781(0.0495) (0.0582) (0.0723) (0.0753) (0.108) (0.112)
Net PP&E/Assets -0.222 -0.207 -0.223 -0.190 0.0368 0.0368(0.179) (0.489) (0.230) (0.336) (0.416) (0.442)
ROA 1.460 1.857 1.410 1.552 3.147 3.147(2.191) (2.510) (2.963) (3.182) (3.518) (3.475)
Intercept -2.324*** -2.325*** -2.766*** -2.718*** -16.59*** -16.59***(0.299) (0.546) (0.424) (0.549) (3.832) (1.205)
Fixed effects None None Cycle Cycle FF-Cycle FF-CycleClustering Firm FF-Cycle Firm FF-Cycle Firm FF-CycleObservations 21,570 21,210 21,570 21,210 14,808 14,808Pseudo-R squared 0.005 0.005 0.236 0.239 0.262 0.262
45
Tab
le3:
Eco
nom
icP
olic
yU
nce
rtai
nty
Sen
siti
vit
yan
dP
olit
ical
Con
trib
uti
on
sT
his
table
docu
men
tsth
eco
rrel
ati
on
bet
wee
nE
conom
icP
olicy
Unce
rtain
tySen
siti
vit
yand
Politi
cal
Contr
ibuti
ons.
The
dep
enden
tva
riable
inco
lum
ns
(1)
–(4
)is
the
natu
ral
logari
thm
of
the
am
ount
afirm
contr
ibute
dto
candid
ate
sin
House
and
Sen
ate
race
s.T
his
vari
able
isth
endec
om
pose
din
Colu
mns
(5)
and
(6)
into
contr
ibuti
ons
toca
ndid
ate
sin
close
elec
tions
(Colu
mn
5)
and
contr
ibuti
ons
toca
ndid
ate
sin
oth
erel
ecti
ons
(Colu
mn
6).
The
dep
enden
tva
riable
inco
lum
ns
(7)
and
(8)
isth
enet
num
ber
of
donati
on
reci
pie
nts
who
won
thei
rre
spec
tive
(clo
se)
elec
tions
wit
hin
agiv
enel
ecti
on
cycl
e.A
llin
dep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.Sta
ndard
erro
rsare
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
Ln(T
ota
lL
n(T
ota
lL
n(T
ota
lL
n(T
ota
lL
n(C
lose
-ele
ctio
nL
n(O
ther
Net
Clo
se-
Net
Clo
se-
Contr
ibuti
ons)
Contr
ibuti
ons)
Contr
ibuti
ons)
Contr
ibuti
ons)
Contr
ibuti
ons)
Contr
ibuti
ons)
Ele
ctio
nW
ins
Ele
ctio
nW
ins
PolicySen
sitive
0.2
07***
0.1
94***
0.0
792**
0.0
749**
0.1
36**
0.0
671*
0.0
657
0.1
02
(0.0
568)
(0.0
293)
(0.0
328)
(0.0
352)
(0.0
555)
(0.0
384)
(0.1
24)
(0.1
36)
Ln
(Size)
0.4
33***
0.4
11***
0.4
37***
0.1
94*
(0.0
438)
(0.0
510)
(0.0
470)
(0.1
04)
BookLev
erage
-0.0
0583
-0.3
01
0.0
936
0.0
963
(0.1
38)
(0.1
93)
(0.1
64)
(0.4
04)
Profit
Marg
in0.0
257
0.0
218
0.0
128
-0.0
611*
(0.0
169)
(0.0
172)
(0.0
246)
(0.0
367)
M/B
0.0
0619*
0.0
0713
0.0
0522
-0.0
029
(0.0
0362)
(0.0
0444)
(0.0
0391)
(0.0
098)
Cash
/Assets
0.0
348
-0.1
45
0.1
72
-0.4
79
(0.1
60)
(0.2
19)
(0.1
75)
(0.4
83)
Intercept
11.1
1***
11.1
1***
12.0
3***
8.2
93***
4.1
24***
8.0
80***
-0.5
40
-2.2
22
(0.0
467)
(0.0
0519)
(0.4
90)
(0.6
14)
(0.8
74)
(0.6
07)
(2.6
35)
(2.8
44)
Fix
edeff
ects
None
Fir
mF
irm
,F
F-C
ycl
eF
irm
,F
F-C
ycl
eF
irm
,F
F-C
ycl
eF
irm
,F
F-C
ycl
eF
irm
,F
F-C
ycl
eF
irm
,F
F-C
ycl
eC
lust
erin
gF
irm
Fir
mF
irm
Fir
mF
irm
Fir
mF
irm
Fir
mO
bse
rvati
ons
27,6
01
27,6
01
27,1
90
23,0
77
23,0
69
22,9
25
27,1
90
23,0
77
R-s
quare
d0.0
03
0.8
57
0.9
05
0.9
10
0.8
13
0.9
02
0.5
15
0.5
31
46
Tab
le4:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
Imp
lied
Vol
atil
ity
This
table
docu
men
tsth
eeff
ects
of
politi
cal
capit
al
shock
son
implied
vola
tility
usi
ng
data
on
firm
s’donati
ons
top
oliti
cal
candid
ate
sin
close
U.S
.fe
der
al
elec
tions.
Panel
Apre
sents
the
diff
eren
ce-i
n-d
iffer
ence
analy
sis
for
firm
sth
at
hav
e“lu
cky”
politi
cal
capit
al
shock
sb
efore
and
aft
eran
elec
tion
com
pare
dto
those
that
had
“unlu
cky”
shock
s.P
anel
Bpre
sents
atr
iple
diff
eren
ceanaly
sis
that
exam
ines
how
this
effec
tva
ries
for
firm
sth
at
are
sensi
tive
toec
onom
icp
olicy
unce
rtain
ty.
Policy
Unce
rtain
tySen
siti
vit
yis
mea
sure
dusi
ng
the
corr
elati
on
bet
wee
na
firm
’seq
uit
yre
turn
sand
the
Baker
,B
loom
,and
Dav
is(2
015)
econom
icp
olicy
unce
rtain
tyin
dex
as
defi
ned
inth
ete
xt.
Daily
implied
vola
tiliti
esare
taken
from
Opti
onM
etri
csand
are
com
pute
dusi
ng
pri
ces
on
at-
the-
money
call
opti
ons.
Each
regre
ssio
nuse
sdaily
data
from
six
month
s’pri
or
toa
feder
al
elec
tion
tosi
xm
onth
sfo
llow
ing
the
elec
tion.
Our
elec
tion
data
spans
all
bie
nnia
lU
.S.
feder
al
elec
tions
from
1998-2
010.
All
indep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.Sta
ndard
erro
rsare
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
Panel
A—
Diff
ere
nce-i
n-D
iffere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
1-M
onth
1-M
onth
3-M
onth
3-M
onth
5-M
onth
5-M
onth
1-M
onth
3-M
onth
5-M
onth
Implied
Implied
Implied
Implied
Implied
Implied
Implied
Implied
Implied
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
PostElection
0.0
349***
0.0
354***
0.0
357***
0.0
360***
0.0
350***
0.0
352***
0.0
172***
0.0
196***
0.0
197***
(0.0
0306)
(0.0
0308)
(0.0
0297)
(0.0
0298)
(0.0
0287)
(0.0
0288)
(0.0
0215)
(0.0
0207)
(0.0
0200)
Post×
Close
Win
Dummy
-0.0
494***
-0.0
495***
-0.0
446***
-0.0
446***
-0.0
421***
-0.0
421***
(0.0
0402)
(0.0
0403)
(0.0
0386)
(0.0
0386)
(0.0
0374)
(0.0
0375)
Post×
Net
Close
Wins
-0.0
117***
-0.0
107***
-0.0
101***
(0.0
00757)
(0.0
00749)
(0.0
00728)
Firm
Return
-0.2
63***
-0.1
76***
-0.1
45***
-0.2
62***
-0.1
76***
-0.1
45***
(0.0
0943)
(0.0
0732)
(0.0
0664)
(0.0
0945)
(0.0
0733)
(0.0
0664)
IndustyReturn
-0.1
62***
-0.0
606***
-0.0
244***
-0.1
63***
-0.0
614***
-0.0
251***
(0.0
126)
(0.0
0953)
(0.0
0852)
(0.0
126)
(0.0
0957)
(0.0
0855)
Intercept
0.3
95***
0.3
95***
0.3
84***
0.3
84***
0.3
77***
0.3
77***
0.3
95***
0.3
84***
0.3
77***
(0.0
0103)
(0.0
0103)
(0.0
00994)
(0.0
00996)
(0.0
00962)
(0.0
00963)
(0.0
0103)
(0.0
00990)
(0.0
00958)
Fix
edeff
ects
Fir
m-C
ycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eC
lust
erin
gF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eO
bse
rvati
ons
841,5
14
841,1
69
839,4
03
839,0
68
830,1
53
829,8
18
841,1
69
839,0
68
829,8
18
R-s
quare
d0.7
48
0.7
50
0.7
88
0.7
89
0.8
02
0.8
03
0.7
51
0.7
90
0.8
04
47
Tab
le4:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
Imp
lied
Vol
atil
ity
(Con
tinu
ed)
Panel
B—
Tri
ple
-Diff
ere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
1-M
onth
1-M
onth
3-M
onth
3-M
onth
5-M
onth
5-M
onth
1-M
onth
3-M
onth
5-M
onth
Implied
Implied
Implied
Implied
Implied
Implied
Implied
Implied
Implied
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
PostElection
0.0
0527**
0.0
0566**
0.0
0599**
0.0
0623**
0.0
0598**
0.0
0617**
-0.0
0503***
-0.0
0245
-0.0
0165
(0.0
0267)
(0.0
0267)
(0.0
0252)
(0.0
0252)
(0.0
0242)
(0.0
0242)
(0.0
0188)
(0.0
0177)
(0.0
0171)
Post×
PolicySen
sitive
0.1
34***
0.1
35***
0.1
35***
0.1
36***
0.1
32***
0.1
32***
0.1
13***
0.1
11***
0.1
07***
(0.0
0889)
(0.0
0894)
(0.0
0869)
(0.0
0873)
(0.0
0843)
(0.0
0846)
(0.0
0731)
(0.0
0713)
(0.0
0686)
Post×
Close
Win
Dummy
-0.0
299***
-0.0
298***
-0.0
244***
-0.0
243***
-0.0
220***
-0.0
219***
(0.0
0351)
(0.0
0351)
(0.0
0329)
(0.0
0329)
(0.0
0320)
(0.0
0320)
Post×
Policy×
Close
Win
Dummy
-0.0
495***
-0.0
499***
-0.0
552***
-0.0
555***
-0.0
572***
-0.0
574***
(0.0
155)
(0.0
156)
(0.0
152)
(0.0
153)
(0.0
146)
(0.0
147)
Post×
Net
Close
Wins
-0.0
0707***
-0.0
0590***
-0.0
0534***
(0.0
00706)
(0.0
00691)
(0.0
00667)
Post×
Policy×
Net
Close
Wins
-0.0
116***
-0.0
129***
-0.0
134***
(0.0
0248)
(0.0
0246)
(0.0
0238)
Firm
Return
-0.2
69***
-0.1
82***
-0.1
51***
-0.2
69***
-0.1
82***
-0.1
51***
(0.0
0947)
(0.0
0733)
(0.0
0659)
(0.0
0949)
(0.0
0734)
(0.0
0660)
IndustyReturn
-0.1
74***
-0.0
727***
-0.0
370***
-0.1
75***
-0.0
734***
-0.0
377***
(0.0
127)
(0.0
0962)
(0.0
0860)
(0.0
127)
(0.0
0963)
(0.0
0861)
Intercept
0.3
95***
0.3
95***
0.3
84***
0.3
84***
0.3
77***
0.3
77***
0.3
95***
0.3
84***
0.3
77***
(0.0
00954)
(0.0
00956)
(0.0
00914)
(0.0
00916)
(0.0
00884)
(0.0
00885)
(0.0
00952)
(0.0
00912)
(0.0
00882)
Fix
edeff
ects
Fir
m-C
ycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eC
lust
erin
gF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eO
bse
rvati
ons
841,5
14
841,1
69
839,4
03
839,0
68
830,1
53
829,8
18
841,1
69
839,0
68
829,8
18
R-s
quare
d0.7
59
0.7
61
0.8
01
0.8
02
0.8
15
0.8
16
0.7
62
0.8
02
0.8
16
48
Tab
le5:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
CD
SS
pre
ads
This
table
docu
men
tsth
eeff
ects
of
politi
cal
capit
al
shock
son
CD
SSpre
ads
usi
ng
data
on
firm
s’donati
ons
top
oliti
cal
candid
ate
sin
close
U.S
.fe
der
al
elec
tions.
Panel
Apre
sents
the
diff
eren
ce-i
n-d
iffer
ence
analy
sis
for
firm
sth
at
hav
e“lu
cky”
politi
cal
capit
al
shock
sb
efore
and
aft
eran
elec
tion
com
pare
dto
those
that
had
“unlu
cky”
shock
s.P
anel
Bpre
sents
atr
iple
diff
eren
ceanaly
sis
that
exam
ines
how
this
effec
tva
ries
for
firm
sth
at
are
sensi
tive
toec
onom
icp
olicy
unce
rtain
ty.
Policy
Unce
rtain
tySen
siti
vit
yis
mea
sure
dusi
ng
the
corr
elati
on
bet
wee
na
firm
’seq
uit
yre
turn
sand
the
Baker
,B
loom
and
Dav
is(2
015)
econom
icp
olicy
unce
rtain
tyin
dex
as
defi
ned
inth
ete
xt.
Daily
CD
Ssp
reads
are
taken
from
Mark
itfo
r1-y
ear,
5-y
ear,
and
10-y
ear
tenors
.A
llC
DS
spre
ads
are
then
expre
ssed
inlo
gfo
rm.
Each
regre
ssio
nuse
sdaily
data
from
six
month
s’pri
or
toa
feder
al
elec
tion
tosi
xm
onth
sfo
llow
ing
the
elec
tion.
Our
elec
tion
data
spans
all
bie
nnia
lU
.S.
feder
al
elec
tions
from
1998-2
010.
All
indep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.Sta
ndard
erro
rsare
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
Panel
A—
Diff
ere
nce-i
n-D
iffere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
1-Y
ear
1-Y
ear
5-Y
ear
5-Y
ear
10-Y
ear
10
Yea
r1-Y
ear
5-Y
ear
10-Y
ear
Log
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SSpre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
PostElection
0.0
439
0.0
781**
0.0
795***
0.0
917***
0.1
05***
0.1
12***
-0.0
672**
0.0
292
0.0
564***
(0.0
322)
(0.0
359)
(0.0
189)
(0.0
207)
(0.0
161)
(0.0
174)
(0.0
336)
(0.0
191)
(0.0
167)
Post×
Close
Win
Dummy
-0.3
09***
-0.3
55***
-0.1
87***
-0.2
10***
-0.1
61***
-0.1
73***
(0.0
393)
(0.0
433)
(0.0
238)
(0.0
254)
(0.0
209)
(0.0
221)
Post×
Close
Wins
-0.0
714***
-0.0
445***
-0.0
367***
(0.0
0695)
(0.0
0444)
(0.0
0388)
Post×
Close
Losses
0.0
767***
0.0
411***
0.0
352***
(0.0
0951)
(0.0
0590)
(0.0
0495)
Firm
Return
0.4
84***
0.3
10***
0.2
41***
0.4
84***
0.3
07***
0.2
40***
(0.0
373)
(0.0
241)
(0.0
238)
(0.0
368)
(0.0
238)
(0.0
236)
Ln
(Size)
-0.7
53***
-0.4
14***
-0.2
49***
-0.7
57***
-0.4
15***
-0.2
50***
(0.1
42)
(0.0
737)
(0.0
637)
(0.1
34)
(0.0
697)
(0.0
614)
M/B
-0.0
0831**
-0.0
0248
-0.0
0496***
-0.0
0840**
-0.0
0251
-0.0
0497***
(0.0
0339)
(0.0
0162)
(0.0
0187)
(0.0
0326)
(0.0
0160)
(0.0
0176)
BookLev
erage
2.9
39***
1.6
72***
1.4
00***
2.7
91***
1.5
94***
1.3
31***
(0.4
34)
(0.2
34)
(0.2
02)
(0.4
17)
(0.2
27)
(0.1
98)
Cash
-0.7
76**
-0.2
13
-0.0
847
-0.7
68**
-0.2
10
-0.0
919
(0.3
70)
(0.2
20)
(0.1
97)
(0.3
65)
(0.2
22)
(0.2
01)
Curren
tRatio
0.0
549
0.0
188
0.0
0271
0.0
517
0.0
176
0.0
0121
(0.0
484)
(0.0
287)
(0.0
235)
(0.0
488)
(0.0
297)
(0.0
242)
Profitability
-1.0
16***
-0.9
00***
-0.5
38**
-0.8
90**
-0.8
23***
-0.4
80**
(0.3
79)
(0.2
50)
(0.2
11)
(0.3
98)
(0.2
46)
(0.2
12)
Intercept
-5.5
23***
-0.2
50
-4.7
47***
-1.9
19***
-4.5
45***
-3.0
89***
-0.1
21
-1.8
67***
-3.0
39***
(0.0
103)
(1.3
22)
(0.0
0618)
(0.6
94)
(0.0
0542)
(0.6
01)
(1.2
53)
(0.6
61)
(0.5
80)
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
355,7
35
249,1
29
388,3
25
271,1
60
359,3
82
251,5
14
249,1
29
271,1
60
251,5
14
R-s
quare
d0.8
98
0.8
95
0.9
20
0.9
26
0.9
17
0.9
22
0.8
96
0.9
27
0.9
23
49
Tab
le5:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
CD
SS
pre
ads
(Con
tinu
ed)
Panel
B—
Tri
ple
-Diff
ere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
1-Y
ear
1-Y
ear
5-Y
ear
5-Y
ear
10-Y
ear
10-Y
ear
1-Y
ear
5-Y
ear
10-Y
ear
Log
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SSpre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
Spre
ad
PostElection
-0.2
35***
-0.2
38***
-0.0
765***
-0.0
822***
-0.0
280*
-0.0
353**
-0.2
57***
-0.0
974***
-0.0
487***
(0.0
299)
(0.0
297)
(0.0
175)
(0.0
173)
(0.0
148)
(0.0
146)
(0.0
209)
(0.0
122)
(0.0
104)
Post×
PolicySen
sitive
0.9
11***
0.9
45***
0.5
17***
0.5
39***
0.4
35***
0.4
49***
0.6
86***
0.3
98***
0.3
42***
(0.0
680)
(0.0
678)
(0.0
414)
(0.0
410)
(0.0
354)
(0.0
352)
(0.0
583)
(0.0
358)
(0.0
307)
Post×
Close
Win
Dummy
-0.0
797**
-0.0
807**
-0.0
667***
-0.0
585***
-0.0
599***
-0.0
528***
(0.0
362)
(0.0
358)
(0.0
217)
(0.0
211)
(0.0
192)
(0.0
186)
Post×
Policy×
Close
Win
Dummy
-0.5
60***
-0.6
07***
-0.2
60***
-0.3
07***
-0.2
00***
-0.2
36***
(0.1
20)
(0.1
25)
(0.0
767)
(0.0
811)
(0.0
660)
(0.0
690)
Post×
Net
Close
Wins
-0.0
258***
-0.0
175***
-0.0
163***
(0.0
0733)
(0.0
0438)
(0.0
0387)
Post×
Policy×
Net
Close
Wins
-0.0
964***
-0.0
573***
-0.0
394***
(0.0
181)
(0.0
117)
(0.0
103)
Firm
Return
0.3
67***
0.2
51***
0.2
00***
0.3
70***
0.2
51***
0.2
01***
(0.0
279)
(0.0
198)
(0.0
193)
(0.0
276)
(0.0
196)
(0.0
192)
M/B
-0.0
00624**
-0.0
00295
-0.0
00189
-0.0
00545*
-0.0
00261
-0.0
00158
(0.0
00281)
(0.0
00189)
(0.0
00180)
(0.0
00283)
(0.0
00192)
(0.0
00181)
Ln
(Size)
-0.4
71***
-0.2
80***
-0.1
51***
-0.4
74***
-0.2
81***
-0.1
52***
(0.1
19)
(0.0
637)
(0.0
570)
(0.1
17)
(0.0
628)
(0.0
567)
BookLev
erage
1.9
54***
1.1
79***
0.9
95***
1.9
43***
1.1
60***
0.9
82***
(0.3
62)
(0.2
02)
(0.1
79)
(0.3
51)
(0.1
96)
(0.1
75)
Profitability
-1.0
84***
-0.8
89***
-0.5
33***
-0.9
25**
-0.7
82***
-0.4
65**
(0.3
44)
(0.2
03)
(0.1
84)
(0.4
09)
(0.2
16)
(0.1
97)
Intercept
-5.5
25***
-2.2
48**
-4.7
50***
-2.8
58***
-4.5
47***
-3.7
96***
-2.2
15**
-2.8
34***
-3.7
75***
(0.0
0916)
(1.1
16)
(0.0
0556)
(0.6
01)
(0.0
0490)
(0.5
40)
(1.1
04)
(0.5
93)
(0.5
35)
Fix
edE
ffec
tF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eC
lust
ered
erro
rsF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eO
bse
rvati
ons
354,1
46
298,3
66
386,5
23
325,0
05
357,7
32
301,9
83
298,3
66
325,0
05
301,9
83
R-s
quare
d0.9
09
0.9
11
0.9
26
0.9
31
0.9
23
0.9
26
0.9
11
0.9
31
0.9
26
50
Tab
le6:
Ter
mStr
uct
ure
Eff
ects
ofP
olit
ical
Cap
ital
Sh
ock
sT
his
table
docu
men
tsth
ete
rmst
ruct
ure
effec
tsof
politi
cal
capit
al
shock
son
CD
Ssp
reads
and
implied
vola
tility
.P
oliti
cal
capit
al
shock
sare
defi
ned
usi
ng
data
on
firm
s’donati
ons
top
oliti
cal
candid
ate
sin
close
U.S
.fe
der
al
elec
tions.
Panel
Apre
sents
the
analy
sis
on
implied
vola
tility
,w
hile
Panel
Bpre
sents
the
analy
sis
on
CD
SSpre
ads.
Inb
oth
panel
s,C
olu
mns
(1)
–(3
)pre
sent
diff
eren
ce-i
n-d
iffer
ence
analy
sis
for
firm
sth
at
hav
e“lu
cky”
politi
cal
capit
al
shock
sb
efore
and
aft
eran
elec
tion
com
pare
dto
those
that
had
“unlu
cky”
shock
s.In
both
Panel
s,C
olu
mns
(4)
–(6
)pre
sent
the
trip
lediff
eren
ceanaly
sis
that
exam
ines
how
this
effec
tva
ries
for
firm
sth
at
are
sensi
tive
toec
onom
icp
olicy
unce
rtain
ty.
Policy
Unce
rtain
tySen
siti
vit
yis
mea
sure
dusi
ng
the
corr
elati
on
bet
wee
na
firm
’seq
uit
yre
turn
sand
the
Baker
,B
loom
and
Dav
is(2
015)
econom
icp
olicy
unce
rtain
tyin
dex
as
defi
ned
inth
ete
xt.
Each
regre
ssio
nuse
sdaily
data
from
six
month
s’pri
or
toa
feder
al
elec
tion
tosi
xm
onth
sfo
llow
ing
the
elec
tion.
Our
elec
tion
data
spans
all
bie
nnia
lU
.S.
feder
al
elec
tions
from
1998-2
010.
Implied
vola
tility
data
isso
urc
edfr
om
Opti
onM
etri
cs.
Daily
CD
Ssp
reads
are
taken
from
Mark
itfo
r1-y
ear,
5-y
ear,
and
10-y
ear
tenors
.Slo
pes
are
defi
ned
as
the
diff
eren
cein
log
CD
Ssp
reads
(colu
mns
(1)-
(6))
or
the
diff
eren
cein
raw
CD
Ssp
reads
(colu
mns
(7)-
(9))
bet
wee
nco
ntr
act
sw
ith
diff
eren
tte
nors
.A
llin
dep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.Sta
ndard
erro
rsare
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
Panel
A—
Implied
Vola
tility
(1)
(2)
(3)
(4)
(5)
(6)
3m
o-
1m
o5m
o-
3m
o5m
o-
1m
o3m
o-
1m
o5m
o-
3m
o5m
o-
1m
oIm
plied
Implied
Implied
Implied
Implied
Implied
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
Vola
tility
PostElection
0.0
00722
-0.0
00633**
0.0
00170
0.0
00788
0.0
00726
2.9
7e-
06
(0.0
00462)
(0.0
00258)
(0.0
00626)
(0.0
00686)
(0.0
00506)
(0.0
00280)
Post×
Close
Win
Dummy
0.0
0445***
0.0
0279***
0.0
0710***
0.0
0759***
0.0
0505***
0.0
0269***
(0.0
00741)
(0.0
00429)
(0.0
0102)
(0.0
0110)
(0.0
00803)
(0.0
00463)
Post×
PolicySen
sitive
-0.0
0281*
-1.9
7e-
05
-0.0
0289***
(0.0
0162)
(0.0
0121)
(0.0
00678)
Post×
Policy×
Close
Win
Dummy
-0.0
0656**
-0.0
0504**
-0.0
0164
(0.0
0273)
(0.0
0200)
(0.0
0115)
Firm
Return
0.0
870***
0.0
320***
0.1
19***
0.1
19***
0.0
870***
0.0
322***
(0.0
0358)
(0.0
0188)
(0.0
0470)
(0.0
0470)
(0.0
0357)
(0.0
0188)
Industry
Return
0.1
01***
0.0
313***
0.1
32***
0.1
32***
0.1
01***
0.0
316***
(0.0
0492)
(0.0
0243)
(0.0
0609)
(0.0
0610)
(0.0
0492)
(0.0
0243)
Intercept
-0.0
114***
-0.0
0628***
-0.0
176***
-0.0
176***
-0.0
114***
-0.0
0628***
(0.0
00177)
(0.0
00101)
(0.0
00242)
(0.0
00241)
(0.0
00177)
(0.0
00101)
Fix
edeff
ects
Fir
m-C
ycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eC
lust
erin
gF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eO
bse
rvati
ons
839,0
68
829,8
18
829,8
18
829,8
18
839,0
68
829,8
18
R-s
quare
d0.1
51
0.1
91
0.2
22
0.2
23
0.1
51
0.1
92
51
Tab
le6:
Ter
mStr
uct
ure
Eff
ects
ofP
olit
ical
Cap
ital
Sh
ock
s(c
onti
nu
ed)
Panel
B—
CD
SSpre
ads
(1)
(2)
(3)
(4)
(5)
(6)
5yr
-1yr
10yr
-5yr
10yr
-1yr
5yr
-1yr
10yr
-5yr
10yr
-1yr
Log
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SL
og
CD
SSpre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
PostElection
0.0
180
0.0
220***
0.0
370*
0.1
51***
0.0
574***
0.2
09***
(0.0
169)
(0.0
0569)
(0.0
217)
(0.0
156)
(0.0
0540)
(0.0
197)
Post×
Close
Win
Dummy
0.1
35***
0.0
402***
0.1
84***
0.0
212
0.0
0350
0.0
304
(0.0
215)
(0.0
0764)
(0.0
275)
(0.0
197)
(0.0
0712)
(0.0
248)
Post×
PolicySen
sitive
-0.3
83***
-0.1
17***
-0.5
06***
(0.0
304)
(0.0
105)
(0.0
389)
Post×
Policy×
Close
Win
Dummy
0.2
71***
0.0
868***
0.3
57***
(0.0
570)
(0.0
208)
(0.0
729)
Firm
Return
-0.1
59***
-0.0
703***
-0.2
38***
-0.1
03***
-0.0
565***
-0.1
64***
(0.0
197)
(0.0
0815)
(0.0
234)
(0.0
142)
(0.0
0677)
(0.0
170)
M/B
0.3
33***
0.1
44***
0.4
64***
0.1
96***
0.1
01***
0.2
84***
(0.0
811)
(0.0
315)
(0.1
11)
(0.0
673)
(0.0
261)
(0.0
913)
Ln
(Size)
0.0
0215*
0.0
0104*
0.0
0315*
0.0
00305***
0.0
00121***
0.0
00424***
(0.0
0119)
(0.0
00533)
(0.0
0169)
(7.0
8e-
05)
(3.4
3e-
05)
(0.0
00103)
BookLev
erage
-1.1
73***
-0.3
63***
-1.5
74***
-0.7
38***
-0.2
18***
-0.9
83***
(0.2
20)
(0.0
824)
(0.2
91)
(0.1
85)
(0.0
703)
(0.2
43)
Profitability
0.2
20
0.1
57*
0.2
98
0.2
85
0.1
63**
0.3
80
(0.2
06)
(0.0
827)
(0.2
74)
(0.2
05)
(0.0
814)
(0.2
72)
Intercept
-1.6
43**
-0.9
23***
-2.4
04**
-0.6
76
-0.6
31***
-1.1
58
(0.7
47)
(0.2
86)
(1.0
15)
(0.6
34)
(0.2
42)
(0.8
50)
Obse
rvati
ons
248,1
67
250,4
70
241,4
01
297,2
60
300,5
48
289,3
25
R-s
quare
d0.7
61
0.8
05
0.8
14
0.7
90
0.8
16
0.8
40
52
Tab
le7:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
Inve
stm
ent,
Cap
ital
Str
uct
ure
,an
dR
&D
Sp
end
ing
Panel
Apre
sents
diff
eren
ces-
in-d
iffer
ence
ste
sts
that
exam
ine
how
“lu
cky”
firm
s’in
ves
tmen
t,le
ver
age,
and
R&
Dsp
endin
gch
ange
follow
ing
ap
osi
tive
shock
toth
efirm
s’p
oliti
cal
capit
al.
Panel
Bpre
sents
atr
iple
-diff
eren
ceanaly
sis
that
exam
ines
how
the
effec
tof
a“lu
cky”
politi
cal
capit
al
shock
vari
esfo
rfirm
sth
at
are
hig
hly
sensi
tive
toec
onom
icp
olicy
unce
rtain
tyw
ithin
agiv
enel
ecti
on
cycl
e.P
olicy
Unce
rtain
tySen
siti
vit
yis
mea
sure
dusi
ng
the
corr
elati
on
bet
wee
na
firm
’seq
uit
yre
turn
sand
the
Baker
,B
loom
and
Dav
is(2
015)
econom
icp
olicy
unce
rtain
tyin
dex
as
defi
ned
inth
ete
xt.
Each
regre
ssio
nis
base
don
quart
erly
data
from
CR
SP
/C
om
pust
at
from
four
quart
ers
pri
or
toea
chel
ecti
on
date
tofo
ur
quart
ers
aft
erea
chel
ecti
on
date
.A
llin
dep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.C
ontr
ol
vari
able
sin
clude
firm
size
,M
/B
,fr
eeca
shflow
/ass
ets,
and
net
PP
&E
/ass
ets
for
inves
tmen
tand
R&
Dre
gre
ssio
ns,
plu
sop
erati
ng
pro
fita
bilit
yand
the
firm
’scu
rren
tra
tio
for
lever
age
regre
ssio
ns.
Banks
and
firm
sm
issi
ng
indust
rycl
ass
ifica
tion
codes
are
excl
uded
from
the
sam
ple
.In
ves
tmen
tand
M/B
are
trim
med
at
the
1%
and
99%
level
s.T
he
sam
ple
isals
ore
stri
cted
tofirm
sw
ith
lever
age
rati
os
bet
wee
nze
roand
one
(incl
usi
ve)
.Sta
ndard
erro
rsare
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
Panel
A—
Diff
ere
nce-i
n-D
iffere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
Inves
tmen
tIn
ves
tmen
tB
ook
Lev
erage
Book
Lev
erage
R&
DSp
endin
gR
&D
Sp
endin
gIt
Kt−
1
It
Assets
t−
1
Lia
bilitie
st
Assets
t
Debt t
Assets
t
R&D
Expenset
Assets
t
R&D
Expenset
Salest
PostElection
0.0
0002
0.0
00194
0.0
0675***
0.0
0114
0.0
00572*
0.0
0322
(0.0
00591)
(0.0
00218)
(0.0
0172)
(0.0
0148)
(0.0
00318)
(0.0
0214)
Post×
Close
Win
Dummy
0.0
0103
0.0
00001
-0.0
0177
-0.0
0162
-0.0
00192
-0.0
0745
(0.0
00880)
(0.0
00320)
(0.0
0239)
(0.0
0235)
(0.0
00510)
(0.0
0579)
Intercept
0.1
56***
0.0
635***
0.5
34***
-0.0
113
0.0
771***
0.7
28**
(0.0
224)
(0.0
0892)
(0.0
788)
(0.0
706)
(0.0
270)
(0.3
10)
Contr
ols
Yes
Yes
Yes
Yes
Yes
Yes
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
18,3
68
18,3
68
18,2
67
18,2
67
6,5
73
6,5
70
R-S
quare
d0.6
91
0.7
83
0.9
47
0.9
41
0.7
21
0.7
52
Panel
B—
Tri
ple
-Diff
ere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
Inves
tmen
tIn
ves
tmen
tB
ook
Lev
erage
Book
Lev
erage
R&
DSp
endin
gR
&D
Sp
endin
g
PostElection
0.0
0210***
0.0
00855***
0.0
0252
-0.0
0246
0.0
00690*
0.0
0430*
(0.0
00674)
(0.0
00244)
(0.0
0183)
(0.0
0167)
(0.0
00403)
(0.0
0240)
Post×
Close
Win
Dummy
-0.0
00622
-0.0
00511
0.0
0183
0.0
0204
-0.0
00146
-0.0
0864
(0.0
00972)
(0.0
00353)
(0.0
0255)
(0.0
0261)
(0.0
00578)
(0.0
0727)
Post×
PolicySen
sitive
-0.0
0844***
-0.0
0268***
0.0
177***
0.0
151***
-0.0
00445
-0.0
0420
(0.0
0139)
(0.0
00536)
(0.0
0438)
(0.0
0375)
(0.0
00826)
(0.0
0651)
Post×
Policy×
Close
Win
Dummy
0.0
0538**
0.0
0168**
-0.0
134**
-0.0
167***
-0.0
00794
0.0
0575
(0.0
0237)
(0.0
00839)
(0.0
0671)
(0.0
0556)
(0.0
0137)
(0.0
118)
Intercept
0.1
64***
0.0
662***
0.5
16***
-0.0
252
0.0
780***
0.7
34**
(0.0
225)
(0.0
0896)
(0.0
781)
(0.0
706)
(0.0
268)
(0.3
09)
Contr
ols
Yes
Yes
Yes
Yes
Yes
Yes
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
18,3
68
18,3
68
18,2
67
18,2
67
6,5
73
6,5
70
R-S
quare
d0.6
93
0.7
84
0.9
48
0.9
41
0.7
21
0.7
52
53
Tab
le7:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
Inve
stm
ent,
Cap
ital
Str
uct
ure
,an
dR
&D
Sp
end
ing
(conti
nu
ed)
Panel
C—
Tri
ple
-Diff
ere
nce
Analy
sis
wit
hC
onti
nuous
Tre
atm
ent
(1)
(2)
(3)
(4)
(5)
(6)
Inves
tmen
tIn
ves
tmen
tB
ook
Lev
erage
Book
Lev
erage
R&
DSp
endin
gR
&D
Sp
endin
g
PostElection
0.0
0187***
0.0
00647***
0.0
0325**
-0.0
0169
0.0
00629**
0.0
0461
(0.0
00500)
(0.0
00175)
(0.0
0144)
(0.0
0125)
(0.0
00268)
(0.0
0287)
Post×
Net
Close
Wins
-0.0
00127
-0.0
00064
0.0
00239
0.0
00378
-0.0
00014
-0.0
00533
(0.0
00185)
(5.8
5e-
05)
(0.0
00508)
(0.0
00536)
(7.7
5e-
05)
(0.0
00629)
Post×
PolicySen
sitive
-0.0
0605***
-0.0
0191***
0.0
120***
0.0
0810***
-0.0
00703
-0.0
0146
(0.0
0114)
(0.0
00404)
(0.0
0334)
(0.0
0282)
(0.0
00663)
(0.0
0410)
Post×
Policy×
Net
Close
Wins
0.0
0148***
0.0
00451**
-0.0
0309**
-0.0
0393***
-0.0
00093
0.0
00077
(0.0
00459)
(0.0
00201)
(0.0
0140)
(0.0
0109)
(0.0
00259)
(0.0
0190)
Intercept
0.1
64***
0.0
662***
0.5
16***
-0.0
257
0.0
780***
0.7
29**
(0.0
225)
(0.0
0893)
(0.0
781)
(0.0
706)
(0.0
269)
(0.3
09)
Contr
ols
Yes
Yes
Yes
Yes
Yes
Yes
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
18,3
68
18,3
68
18,2
67
18,2
67
6,5
73
6,5
70
R-S
quare
d0.6
93
0.7
84
0.9
48
0.9
41
0.7
21
0.7
52
54
Tab
le8:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
Op
erat
ing
Per
form
ance
and
Pro
fita
bil
ity
Panel
Apre
sents
diff
eren
ces-
in-d
iffer
ence
ste
sts
that
exam
ine
how
“lu
cky”
firm
s’op
erati
ng
per
form
ance
and
pro
fita
bilit
ych
ange
follow
ing
ap
osi
tive
shock
toth
efirm
s’p
oliti
cal
capit
al.
Panel
Bpre
sents
atr
iple
-diff
eren
ceanaly
sis
that
exam
ines
how
the
effec
tof
a“lu
cky”
politi
cal
capit
al
shock
vari
esfo
rfirm
sth
at
are
hig
hly
sensi
tive
toec
onom
icp
olicy
unce
rtain
tyw
ithin
agiv
enel
ecti
on
cycl
e.P
olicy
Unce
rtain
tySen
siti
vit
yis
mea
sure
dusi
ng
the
corr
elati
on
bet
wee
na
firm
’seq
uit
yre
turn
sand
the
Baker
,B
loom
and
Dav
is(2
015)
econom
icp
olicy
unce
rtain
tyin
dex
as
defi
ned
inth
ete
xt.
Each
regre
ssio
nis
base
don
quart
erly
data
from
CR
SP
/C
om
pust
at
from
four
quart
ers
pri
or
toea
chel
ecti
on
date
tofo
ur
quart
ers
aft
erea
chel
ecti
on
date
.A
llin
dep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.B
anks
and
firm
sm
issi
ng
indust
rycl
ass
ifica
tion
codes
are
excl
uded
from
the
sam
ple
.M
/B
istr
imm
edat
the
1%
and
99%
level
s.Sta
ndard
erro
rsare
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10%
,5%
,and
1%
level
s,re
spec
tivel
y.
Panel
A—
Diff
ere
nce-i
n-D
iffere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Sale
sA
sset
sM
/B
RO
AC
OG
SSG
&A
Pro
fit
Marg
in
ln(S
ales t
)ln
(Assets t
)M
/B
tEBITt
Assets
t−
1
COGSt
Salest
SG&A
tSalest
EBITD
At
Salest
PostElection
0.0
402***
0.0
743***
-0.2
13***
-0.0
0243***
-0.0
00225
-0.0
0629
0.0
0556
(0.0
0567)
(0.0
0532)
(0.0
387)
(0.0
00366)
(0.0
152)
(0.0
0778)
(0.0
230)
Post×
Close
Win
Dummy
0.0
198**
-0.0
0955
0.1
69***
0.0
0238***
0.0
0389
-0.0
0222
-0.0
0440
(0.0
0880)
(0.0
0783)
(0.0
557)
(0.0
00521)
(0.0
165)
(0.0
0952)
(0.0
244)
Intercept
7.1
46***
8.8
64***
2.9
53***
0.0
234***
0.6
88***
0.2
16***
0.0
994***
(0.0
0216)
(0.0
0195)
(0.0
140)
(0.0
00132)
(0.0
0462)
(0.0
0253)
(0.0
0692)
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
22,2
96
22,3
53
21,1
52
21,9
13
22,1
90
15,8
37
22,1
84
R-S
quare
d0.9
84
0.9
93
0.8
57
0.6
95
0.6
07
0.7
46
0.6
79
Panel
B—
Tri
ple
-Diff
ere
nce
Analy
sis
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Sale
sA
sset
sM
/B
RO
AC
OG
SSG
&A
Pro
fit
Marg
in
PostElection
0.0
699***
0.0
885***
-0.0
953**
-0.0
0158***
-0.0
0753
-0.0
0971
0.0
184
(0.0
0592)
(0.0
0625)
(0.0
403)
(0.0
00380)
(0.0
194)
(0.0
102)
(0.0
295)
Post×
Close
Win
Dummy
-0.0
0697
-0.0
214**
0.0
692
0.0
0166***
0.0
124
0.0
00478
-0.0
172
(0.0
0947)
(0.0
0882)
(0.0
589)
(0.0
00557)
(0.0
207)
(0.0
119)
(0.0
310)
Post×
PolicySen
sitive
-0.1
32***
-0.0
628***
-0.5
25***
-0.0
0379***
0.0
323
0.0
141
-0.0
568*
(0.0
146)
(0.0
112)
(0.1
06)
(0.0
0102)
(0.0
222)
(0.0
105)
(0.0
316)
Post×
Policy
×Close
Win
Dummy
0.1
06***
0.0
421**
0.3
65**
0.0
0265**
-0.0
427*
-0.0
0717
0.0
563*
(0.0
214)
(0.0
188)
(0.1
59)
(0.0
0134)
(0.0
239)
(0.0
125)
(0.0
333)
Intercept
7.1
46***
8.8
64***
2.9
53***
0.0
234***
0.6
88***
0.2
16***
0.0
994***
(0.0
0212)
(0.0
0194)
(0.0
139)
(0.0
00131)
(0.0
0462)
(0.0
0253)
(0.0
0692)
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
22,2
96
22,3
53
21,1
52
21,9
13
22,1
90
15,8
37
22,1
84
R-S
quare
d0.9
84
0.9
93
0.8
58
0.6
96
0.6
07
0.7
46
0.6
79
55
Tab
le8:
Th
eIm
pac
tof
Pol
itic
alC
apit
alS
hock
son
Op
erat
ing
Per
form
ance
and
Pro
fita
bil
ity
(conti
nu
ed)
Panel
C—
Tri
ple
-Diff
ere
nce
Analy
sis
wit
hC
onti
nuous
Tre
atm
ent
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Sale
sA
sset
sM
/B
RO
AC
OG
SSG
&A
Pro
fit
Marg
in
PostElection
0.0
666***
0.0
803***
-0.0
654**
-0.0
00933***
-0.0
0167
-0.0
0946
0.0
107
(0.0
0469)
(0.0
0452)
(0.0
303)
(0.0
00289)
(0.0
114)
(0.0
0626)
(0.0
172)
Post×
Net
Close
Wins
0.0
00716
-0.0
0397**
0.0
0308
0.0
00282***
-0.0
00928
-7.7
5e-
05
0.0
00167
(0.0
0187)
(0.0
0166)
(0.0
113)
(0.0
00105)
(0.0
0146)
(0.0
00600)
(0.0
0162)
Post×
PolicySen
sitive
-0.0
839***
-0.0
446***
-0.3
70***
-0.0
0235***
0.0
118
0.0
103
-0.0
317*
(0.0
113)
(0.0
0918)
(0.0
803)
(0.0
00692)
(0.0
133)
(0.0
0655)
(0.0
185)
Post×
Policy
×Net
Close
Wins
0.0
270***
0.0
108***
0.1
04***
0.0
0114***
-0.0
0900*
-0.0
0248**
0.0
105**
(0.0
0463)
(0.0
0352)
(0.0
371)
(0.0
00378)
(0.0
0495)
(0.0
0112)
(0.0
0493)
Intercept
7.1
46***
8.8
64***
2.9
53***
0.0
234***
0.6
88***
0.2
16***
0.0
995***
(0.0
0211)
(0.0
0194)
(0.0
139)
(0.0
00131)
(0.0
0461)
(0.0
0253)
(0.0
0691)
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
22,2
96
22,3
53
21,1
52
21,9
13
22,1
90
15,8
37
22,1
84
R-S
quare
d0.9
84
0.9
93
0.8
58
0.6
96
0.6
07
0.7
46
0.6
79
56
Table 9: Political Capital Shocks and Marginal Expected ShortfallThe following table presents regression estimates of various connection measures on firm Marginal Expected Shortfall(MES). MES, which is defined in Acharya, Pedersen, Philippon, and Richardson (2010), represents the conditionalexpected return on a stock given a left-tail market return realization. A more negative MES number indicates agreater exposure to systemic/tail risk. All independent variables are defined in the Appendix. All regressions includefirm-election cycle fixed effects. Standard errors clustered by firm are reported in parentheses. *, **, and *** denotestatistical significance at the 10, 5, and 1 % levels respectively.
Political Capital Shocks and MES(1) (2) (3) (4) (5) (6)
Marginal Marginal Marginal Marginal Marginal MarginalExpected Expected Expected Expected Expected ExpectedShortfall Shortfall Shortfall Shortfall Shortfall Shortfall
PostElection -0.00174*** -0.00134*** -0.00250*** -0.00189*** -0.00102*** -0.000817***(0.000206) (0.000168) (0.000275) (0.000224) (0.000299) (0.000238)
Post×Net Close Wins 0.000566*** 0.000446***(9.86e-05) (8.07e-05)
Post× Close Win Dummy 0.00205*** 0.00150***(0.000457) (0.000365)
Post× Close Wins 0.000506*** 0.000402***(9.67e-05) (7.99e-05)
Post× Close Losses -0.000767*** -0.000592***(0.000119) (9.42e-05)
MarketCap 4.36e-05 -0.00203** 0.000118 -0.00200** 8.45e-05 -0.00202**(0.000845) (0.000827) (0.000845) (0.000828) (0.000845) (0.000825)
CAPM Beta -0.0231*** -0.0231*** -0.0230***(0.000699) (0.000700) (0.000698)
CAPM V olatility -0.000274 -0.000305 -0.000295(0.000320) (0.000320) (0.000320)
Intercept -0.0270** 0.0271** -0.0281** 0.0266** -0.0276** 0.0268**(0.0129) (0.0130) (0.0129) (0.0130) (0.0129) (0.0130)
Fixed effects Firm-cycle Firm-cycle Firm-cycle Firm-cycle Firm-cycle Firm-cycleClustered errors Firm Firm Firm Firm Firm FirmObservations 7,792 7,720 7,792 7,720 7,792 7,720R-squared 0.881 0.924 0.881 0.924 0.882 0.925
57
Tab
le10
:G
over
nm
ent
Con
trac
tors
and
CD
SS
pre
ads
The
follow
ing
table
pre
sents
regre
ssio
nes
tim
ate
sofva
rious
politi
calco
nnec
tion
mea
sure
son
log
CD
Ssp
reads
for
“gov
ernm
ent-
dep
enden
t”ver
sus
“non-g
over
nm
ent-
dep
enden
t”firm
s.Sp
ecifi
cati
ons
(1)
-(3
)are
run
on
the
full
sam
ple
of
firm
s.Sp
ecifi
cati
ons
(4)
-(6
)are
run
on
the
sam
ple
of
Gov
ernm
ent
Supplier
s(s
eete
xt
for
det
ails)
,w
hile
spec
ifica
tions
(7)
-(9
)ex
clude
gov
ernm
ent
supplier
s.A
llin
dep
enden
tva
riable
sare
defi
ned
inth
eA
pp
endix
.A
llre
gre
ssio
ns
incl
ude
firm
-ele
ctio
ncy
cle
fixed
effec
ts.
Sta
ndard
erro
rscl
ust
ered
by
firm
-ele
ctio
ncy
cle
are
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10,
5,
and
1%
level
sre
spec
tivel
y.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
(9)
Sam
ple
:A
llA
llA
llG
ovG
ovG
ovN
on-G
ovN
on-G
ovN
on-G
ovF
irm
sF
irm
sF
irm
sF
irm
sF
irm
sF
irm
sF
irm
sF
irm
sF
irm
s
Dep
enden
tV
ari
able
:1-Y
ear
5-Y
ear
10-Y
ear
1-Y
ear
5-Y
ear
10-Y
ear
1-Y
ear
5-Y
ear
10-Y
ear
Spre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
Spre
ads
PostElection
0.0
22
0.0
72***
0.1
00***
0.0
45
0.0
84
0.0
87
0.0
36*
0.0
88***
0.1
09***
(0.0
27)
(0.0
16)
(0.0
14)
(0.2
54)
(0.1
64)
(0.1
43)
(0.0
33)
(0.0
20)
(0.1
76)
Gov
Con
tractor
-0.2
09
-0.1
85
-0.1
01
(0.1
93)
(0.1
16)
(0.0
99)
Post×
Close
ElectionDummy
-0.3
02***
-0.1
86***
-0.1
61***
-0.2
40
-0.2
27
-0.2
76*
-0.3
38***
-0.2
02***
-0.1
70***
(0.0
36)
(0.0
22)
(0.0
19)
(0.3
08)
(0.2
08)
(0.1
65)
(0.0
40)
(0.0
24)
(0.0
21)
Post×
Gov
Con
tractor
0.1
65
0.1
33
0.0
464
(0.2
04)
(0.1
22)
(0.1
05)
Gov
×Close
ElectionDummy
0.0
695
0.1
39
0.1
17
(0.2
13)
(0.1
29)
(0.1
09)
Post×
Close
ElectionDummy×
Gov
-0.1
93
-0.1
77
-0.1
29
(0.2
28)
(0.1
38)
(0.1
16)
Intercept
-5.5
07***
-4.7
30***
-4.5
35***
-5.5
16***
-4.6
19***
-4.4
10***
-5.5
99***
-4.8
09***
-4.6
04***
(0.0
10)
(0.0
06)
(0.0
05)
(0.1
54)
(0.1
02)
(0.0
857)
(0.0
10)
(0.0
06)
(0.0
05)
Fix
edeff
ects
Fir
m-c
ycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eC
lust
ered
erro
rsF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eF
irm
-cycl
eO
bse
rvati
ons
395,9
68
438,6
40
401,2
02
9,4
52
10,3
27
9,5
97
304,9
38
333,5
15
308,6
05
R-s
quare
d0.8
91
0.9
18
0.9
15
0.9
74
0.9
81
0.9
82
0.8
91
0.9
15
0.9
10
58
Tab
le11
:In
du
stry
Rob
ust
nes
sT
ests
This
table
exam
ines
the
robust
nes
sof
the
test
sin
Table
s4
-8
todiff
eren
tty
pes
of
indust
ryco
ntr
ols
.Sp
ecifi
cally,
the
regre
ssio
ns
bel
owall
incl
ude
firm
and
indust
ry-c
ycl
efixed
effec
ts.
As
such
,th
ese
test
slo
okwithin
an
indust
rywithin
agiv
enel
ecti
on
cycl
efo
rid
enti
fica
tion.
Sta
ndard
erro
rscl
ust
ered
by
firm
-ele
ctio
ncy
cle
are
rep
ort
edin
pare
nth
eses
.T
he
sym
bols
*,
**,
and
***
den
ote
stati
stic
al
signifi
cance
at
the
10,
5,
and
1%
level
sre
spec
tivel
y.
(1)
(2)
(3)
(4)
(5)
(6)
(7)
(8)
5-y
ear
3-m
onth
Inves
tmen
tL
ever
age
M/B
Sale
sA
sset
sR
OA
CD
SIm
plied
Spre
ads
Vola
tility
PostElection
-0.0
761***
-0.0
037**
0.0
00679
0.0
00689
-0.0
706*
0.0
672***
0.0
868***
-0.0
014***
(0.0
176)
(0.0
018)
(0.0
00623)
(0.0
01687)
(0.0
396)
(0.0
058)
(0.0
061)
(0.0
004)
Close
Win
Dummy
0.0
00972
0.0
030***
-0.0
01775*
0.0
03993
-0.1
625**
0.0
002
0.0
456***
-0.0
022***
(0.0
334)
(0.0
009)
(0.0
00962)
(0.0
03593)
(0.0
762)
(0.0
160)
(0.0
153)
(0.0
006)
PolicySen
sitiveDummy
-0.2
49***
-0.0
555***
0.0
04516***
-0.0
03181
0.2
481**
0.0
553**
0.0
354
0.0
015*
(0.0
439)
(0.0
064)
(0.0
01530)
(0.0
05225)
(0.1
259)
(0.0
218)
(0.0
228)
(0.0
009)
Post×
Close
Win
Dummy
-0.0
634***
-0.0
060***
-0.0
00126
0.0
02555
0.0
374
-0.0
053
-0.0
190**
0.0
014***
(0.0
218)
(0.0
007)
(0.0
00886)
(0.0
02551)
(0.0
577)
(0.0
091)
(0.0
086)
(0.0
005)
Post×
PolicySen
sitiveDummy
0.5
15***
0.1
099***
-0.0
07280***
0.0
09918***
-0.4
974***
-0.1
299***
-0.0
644***
-0.0
039***
(0.0
414)
(0.0
071)
(0.0
01258)
(0.0
03425)
(0.1
017)
(0.0
140)
(0.0
109)
(0.0
010)
Policy×
Close
Win
Dummy
0.0
393
0.0
047**
-0.0
02455
0.0
05116
0.1
212
-0.0
183
-0.0
356
0.0
01
(0.0
742)
(0.0
022)
(0.0
02456)
(0.0
08340)
(0.2
101)
(0.0
330)
(0.0
346)
(0.0
014)
Post×
Policy
×Close
Win
Dummy
-0.2
67***
-0.0
123***
0.0
05228**
-0.0
11221**
0.4
318***
0.0
941***
0.0
338*
0.0
025*
(0.0
772)
(0.0
024)
(0.0
02232)
(0.0
05387)
(0.1
573)
(0.0
207)
(0.0
190)
(0.0
013)
Intercept
-4.7
06***
0.3
939***
0.0
74579***
0.2
13552***
0.2
173
6.8
751***
8.0
019***
0.0
395***
(0.0
196)
(0.0
459)
(0.0
14451)
(0.0
68447)
(1.1
879)
(0.4
740)
(0.3
430)
(0.0
122)
Fix
edeff
ects
Fir
m,
Fir
m,
Fir
m,
Fir
m,
Fir
m,
Fir
m,
Fir
m,
Fir
m,
FF
-Cycl
eF
F-C
ycl
eF
F-C
ycl
eF
F-C
ycl
eF
F-C
ycl
eF
F-C
ycl
eF
F-C
ycl
eF
F-C
ycl
eC
lust
ered
erro
rsF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eF
irm
-Cycl
eO
bse
rvati
ons
379,5
03
828508
20458
22353
21152
22296
22353
21913
R-S
quare
d0.8
53
0.6
93
0.5
26
0.8
48
0.6
85
0.9
58
0.9
72
0.5
51
59