Examine the determinants and consequences of investor conferences.

Contents lists available at ScienceDirect
Journal of Accounting and Economics
journal homepage: www.elsevier.com/locate/jae
Corporate jets and private meetings with investors☆
Brian J. Busheea, Joseph Gerakosb, Lian Fen Leec,⁎
a The Wharton School, University of Pennsylvania, 1300 Steinberg-Dietrich Hall, Philadelphia, PA 19104-6365, United States
b Tuck School of Business, Dartmouth College, 100 Tuck Hall, Hanover, NH 03755, United States
c Carroll School of Management, Boston College, 140 Commonwealth Ave., Chestnut Hill, MA 02467, United States
A R T I C L E I N F O
Keywords:
Selective disclosure
Corporate jets
Institutional investors
JEL classification:
G14
K22
M48
A B S T R A C T
We use corporate jet flight patterns to identify private meetings with investors that are ex ante
unobservable to non-participants. Using approximately 400,000 flights, we proxy for private
meetings with “roadshows,” defined as three-day windows that include flights to money centers
and to non-money centers in which the firm has high institutional ownership. Roadshows exhibit
greater abnormal stock reactions, analyst forecast activity, and absolute changes in local institutional
ownership than other flight activity. We also find positive trading gains in firms with
more complex information and infrequent private meetings, suggesting that roadshows provide
participating investors an advantage over non-participating investors.
1. Introduction
We examine whether corporate jet flight patterns can be used to identify private meetings between managers and investors that
are ex ante unobservable to non-participants. Such private meetings potentially provide information to participants even if managers
do not directly disclose new material information. For example, participants who devote resources to gathering and processing
information about the firm have the opportunity to refine their private information, which is allowed under Regulation Fair
Disclosure (Reg FD). While unobservable private interactions between managers and investors could take place through a number of
media, such as phone calls and emails, there are no good proxies for such communications. Our corporate jet sample allows us to
plausibly identify such unobservable private meetings, and hence measure their impact on stock returns, analyst forecasts, and
institutional investor trading.
Recent research provides evidence of private meetings between managers and investors at investor conferences and analyst/
investor days (see Koch et al., 2013 for a partial review). Although these meetings are private, they occur at public events that are
scheduled well in advance and, thus, non-participants likely know that these meetings occur. By contrast, non-participants are likely
unaware of private meetings facilitated by corporate jet flights and are therefore unable to modify their trading behavior in anticipation
of the meetings (e.g., stay out of the market). Consequently, these private meetings could place non-participants at a
significant trading disadvantage because they do not know that some investors’ trades are informed by the meetings. We assess
whether these unobservable private meetings put some investors at an informational disadvantage by providing evidence on whether
https://doi.org/10.1016/j.jacceco.2018.01.005
Received 9 January 2015; Received in revised form 19 January 2018; Accepted 30 January 2018
☆ The authors are grateful for financial support provided by the Carroll School of Management, the Tuck School of Business, and The Wharton School. We thank
Gavin Seng for programming assistance. We thank Dan Cohen, Joao Granja, Frank Hodge, Xiumin Martin, Dawn Matsumoto, Eddie Riedl, an anonymous reviewer,
Eugene Soltes, Suraj Srinivasan, Joanna Wu (the editor), and workshop participants at Arizona State University, Boston College, Boston University, Carnegie Mellon
University, Harvard Business School, INSEAD, London Business School, the London School of Economics, MIT, the 2012 Notre Dame Accounting Conference, the Ohio
State University, Santa Clara University, Singapore Management University, the 2013 Stanford Summer Accounting Camp, Tilburg University, University of
Connecticut, University of Kansas, University of Missouri, University of Washington, and Washington University at St. Louis for their comments.
⁎ Corresponding author.
E-mail addresses: bushee@wharton.upenn.edu (B.J. Bushee), joseph.j.gerakos@tuck.dartmouth.edu (J. Gerakos), lianfen.lee@bc.edu (L.F. Lee).
Journal of Accounting and Economics 65 (2018) 358–379
Available online 01 February 2018
0165-4101/ © 2018 Elsevier B.V. All rights reserved.
T
these meetings are associated with stock market reactions and analyst forecast activity during the flight windows, as well as with
changes in local institutional ownership and local investor trading gains.
Corporate jet flights are unobservable because companies routinely request that the Federal Aviation Administration block their
flight plans from real-time tracking due to privacy concerns, indicating that managers do not want their jet usage tracked
(Maremont and McGinty, 2011). We collect corporate jet flight patterns from the Wall Street Journal Jet Tracker database. The Wall
Street Journal filed a Freedom of Information Act request in 2011 to get historical access to flight plans between 2007 and 2010. Our
data consists of almost 400,000 flights undertaken by 396 firms.
We use the flight data to form non-overlapping three-day flight windows. The data does not allow us to identify the passengers on
a given flight or the purpose of the trip. Instead, we use flight patterns to construct a measure called “roadshows” to proxy for
unobservable private meetings between managers and investors. We define a “money center roadshow” (MC Roadshow) as a threeday
flight window in which there are flights to two or more money centers, i.e., Boston, Chicago, New York, and San Francisco
(Bushee et al., 2011). Given the high concentration of current and potential investors located in those cities, flights to money centers
are more likely to be associated with investor meetings than flights to other locations. Moreover, a MC Roadshow flight window is
more likely to be for investor meetings than a flight window with flights to only one money center (denoted as MC Single), which
could be for operational or personal reasons.
We recognize that there may be other reasons why managers fly to multiple money centers in a flight window. Thus, we also
define a “non-money-center” roadshow (Non-MC Roadshow) as a three-day flight window during which there are flights to two or
more “high ownership” non-money-center cities. For each firm, we identify “high-ownership” cities that are not money centers based
on the number and percent of firm-specific institutional ownership in the area. These high-ownership cities differ across firms and are
thus less likely to represent the effect of some other economic event that is specific to money centers. We expect that Non-MC
Roadshow is more likely to proxy for investor meetings than a flight window with flights to only one high-ownership city (Non-MC
Single) or to no high-ownership cities (denoted as Other).
To mitigate the possibility that roadshow flight windows still contain private meetings that occur at the same time as a public
information event, we remove any flight windows with days that contain SEC filings (10-K, 10-Q, 8-K), earnings announcements, and
conference presentations. We also search Factiva for any public notice of meetings for the 500 flight windows with the largest stock
return reactions and remove any observations with a public news event. Thus, the roadshows are likely to capture unobservable
private meetings.
We first validate that our roadshow variables proxy for unobservable private meetings by examining the determinants of roadshow
flights. Consistent with incentives for a manager to incur the cost of flying out to meet key investors, the number of money
center roadshows in a month is positively associated with the proportion of the firm’s intangible assets (a proxy for information
complexity), the level of institutional investor ownership (a proxy for investor demands), events that can change the demand for
information about the firm (i.e., upcoming debt/equity issue and a recent earnings announcement), and a 12-month lagged measure
of roadshows (a proxy for a commitment to provide transparency to selected key investors). For non-money-center roadshows, we
find that institutional investor ownership, an upcoming debt/equity issue, and the 12-month lagged measure of roadshows are
significant determinants of the number of roadshows. When we examine the determinants of other flights (i.e., not roadshows), none
of our determinants has a significant coefficient with the same sign as in the roadshow regressions. Thus, our results suggest that
roadshow flights reflect incentives to meet with investors, rather than general motivations for a high number of flights in a month.
Next, we examine the stock market reaction during the roadshow flight windows. Because of Reg FD, managers are unlikely to
disclose material quantitative information (e.g., earnings forecasts) during private meetings with investors. Rather, any market
reaction would likely stem from investors revising their beliefs based on qualitative information that complements their private
information, which is consistent with the “mosaic” view permitted under Reg FD (Cooley Godward, 2000). Because the direction of
such revisions is unclear ex ante in this setting, we measure the stock market reaction with abnormal absolute size-adjusted returns
and abnormal share turnover.
We find significant positive abnormal market reactions to MC Roadshow and Non-MC Roadshow flight windows that are significantly
greater than market reactions to MC Single and Non-MC Single windows. The coefficients suggest a 4.0% (3.6%) relative
increase in absolute size-adjusted returns (share turnover) for money-center roadshows compared to other flight windows. The
economic magnitudes are relatively small because private meetings are measured with noise and only a small subset of investors meet
with managers at these meetings.1 However, the fact that we statistically detect a market reaction due to a small subset of investors
updating their private information (while other investors are unaware of the meetings) suggests that those investors believe that they
receive an information advantage through the meetings.
In addition, we find positive signed size-adjusted returns during both types of roadshows. These findings suggest that managers
are more likely to undertake roadshows when they believe the stock price is undervalued. This evidence is also more consistent with
the investor meetings triggering the price reaction than the alternative story that initial price movement leads managers to fly to
investors to manage the news, which would be more likely in the case of bad news.
While our main focus is investor meetings, we also examine whether roadshows affect security analysts. If managers use corporate
jets to also meet with analysts, we expect MC Roadshow to be associated with analyst forecast activity because of the high
1 The market reaction to a roadshow is not a measure of the information content of the meetings, as would be the case with a public disclosure by management.
While the investors privately meeting with managers would incorporate any information into their trades, other investors are likely unaware of the meetings and could
therefore trade to “correct” what appears to be an unwarranted market reaction. Such trades would attenuate the magnitude of abnormal returns.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
359
concentration of analysts located in money centers. By contrast, we do not expect Non-MC Roadshow to be a reasonable proxy for
private meetings with analysts because these windows are defined based on high institutional ownership, not analyst following. Thus,
finding differences between MC Roadshow and Non-MC Roadshow would further validate that our results are driven by private
meetings, rather than some more general phenomenon. We find that the number of analysts who issue forecasts during a flight
window is positively associated with MC Roadshow, but not associated with Non-MC Roadshow.
Next, we examine whether roadshow flights to a given Metropolitan Statistical Area (MSA) are associated with quarterly changes
in ownership by institutional investors located in that MSA. We examine both signed and unsigned changes in local institutional
ownership because it is ex ante unclear which direction the meetings will update investors’ beliefs about the firm. We find that MC
Roadshow is associated with an increase in local institutional ownership, and that both MC Roadshow and Non-MC Roadshow are
associated with significant absolute changes in local ownership. Thus, roadshows are associated with more buying and selling by
local institutional investors in the firm’s stock.
While we find that local institutional investors trade in response to the roadshows, it is not clear whether such trading is informed.
Thus, we investigate the potential trading gains of local institutional investors after the roadshows. We estimate trading gains by
multiplying the quarterly changes in local institutional investor holdings by the firm’s buy-and-hold size-adjusted returns over the
subsequent quarter; thus, trading gains are positive when institutional investors trade in the correct direction of future returns. We
find that roadshows are not unconditionally related to greater trading gains. However, there are significant positive trading gains for
MC Roadshow for firms with more complex information (high intangible assets) and where private interactions between managers
and investors are previously infrequent (prior flights to the MSA are lower). Thus, while there is no widespread evidence of institutional
investors being able to earn trading gains based on roadshow meetings, the roadshows do appear beneficial to investors in
certain limited contexts.
Overall, we document an important channel of private firm communication with institutional investors that has not been explored
in prior research. Unobservable private communications between managers and investors can occur through a number of media, such
as phone calls and emails, but it is empirically challenging to capture such communications. Prior work finds significant market
reactions to settings that involve private meetings (e.g., Bushee et al., 2011,2017; Green et al., 2014a, b; Soltes, 2014; Kirk and
Markov, 2016; Cheng et al., 2016). However, these papers examine private meetings at well-publicized investor conferences and
analyst/investor days. Solomon and Soltes (2015) study all private meetings with investors for one firm, but their sample is a mix of
meetings at conferences (two-thirds of their sample) and unobservable private meetings (one-third). Our setting is unique because the
presence of all of these flights is not publicly-known, thereby allowing us to examine unobservable private meetings across a large
sample of firms.2
Because of the absence of publicly-available information on the corporate-jet-facilitated private meetings, excluded investors
would have to engage in high-cost behavior to know of their existence. For example, an investor could station people at airports to
“tail spot” private jets, and thus learn that managers are flying to a given city. However, only those investors willing to incur such a
high cost would be aware that meetings could be occurring, and it is likely that most investors would still be unaware of the meetings.
In fact, managers continue to block public release of their flight plans, even after the one-time release of data to the Wall Street
Journal.3 By contrast, it is fairly low cost for excluded investors to learn about private meetings that occur during publicly-announced
investor conferences in advance through Thomson Reuters, StreetEvents, Bloomberg and even the companies themselves. Thus, from
an ex ante perspective, the substantially higher costs of detecting private meetings for excluded investors in our sample increases the
probability that these private meetings occur in a setting in which non-participants likely do not know that they are at a potential
information disadvantage.
While one could argue that the findings in prior research could extrapolate to our setting, the policy implications from the findings
differ. Regulators are less likely to be concerned about private meetings if non-participants are aware of the occurrence of such
meetings and could potentially stay out of the market. By contrast, our study suggests greater potential harm to non-participants who
are unaware of their informational disadvantage. Thus, our results should appeal to regulators interested in leveling the playing field
for all investors. Our findings should also be of interest to investors who are potentially hurt by non-participation in these meetings,
and those interested in the extent and nature of firms’ private meetings with selected investors.
Our findings also relate to the academic literature on corporate jets. Prior research on corporate jets often focuses on potential
agency costs. Several papers provide evidence consistent with managers using corporate jets for private benefits at the expense of firm
value (e.g. Yermack, 2006; Andrews et al., 2009; Edgerton, 2011 and Grinstein et al., 2017) while other studies suggest that use of
corporate jets could increase productivity and efficiency, and thus, is a rational business expense (e.g. Rajan and Wulf, 2006; Lee
et al., 2018). Our findings provide some insights into how managers use corporate jets for roadshows to communicate privately with
selected investors. They do not, however, speak directly to whether corporate jets are value increasing.
2 The difference in the extent to which the private meetings in our setting contain public information events compared to settings in prior literature could explain the
smaller economic magnitudes of the market reaction to private meetings in our study. For example, Bushee et al. (2011) reports a 9% increase in absolute SAR during
conference presentation windows, which involve a public presentation and greater investor participation. They also report that, for Form 8-K filings (earnings
announcements), the absolute SAR increases by 31% (84%).
3 We searched for real-time flight plans for a number of companies and, in each case, received a message that “this aircraft is not available for live tracking.”
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
360
2. Motivation and predictions
2.1. Prior research on direct communications with analysts and investors
The literature on face-to-face management communications with investors and analysts motivates our examination of managers’
use of corporate jets for roadshows. In October 2000, the SEC enacted Reg FD to stop the practice of selective disclosure of material
information by managers to investors and analysts. Numerous studies indicate that Reg FD “leveled the playing field” by providing
individual investors greater real-time access to disclosure and by reducing analysts’ private access to information (see Koch et al.,
2013 for a review). However, Reg FD does not explicitly preclude private meetings. Managers have latitude to discuss details of the
business to fill in the mosaic of investors’ or analysts’ private information without violating Reg FD (Cooley Godward, 2000).4
In our context, we do not claim that managers disclose material information or intend to convey a specific piece of information to
investors when they go on roadshows. It is unlikely that they would do so given concerns about violating Reg FD. Even if this were the
case and all participants receive identical information, research suggests that no investor is expected to profit (Sidhu et al., 2008). We
argue it is more likely that participants at these private meetings have the opportunity to gather qualitative information that supplements
their private information. Thus, investors who invest resources gathering and processing information before the meetings
are better prepared to ask more insightful questions, which enable them to confirm or update their beliefs.
Recent research suggests that private meetings are prevalent subsequent to Reg FD. Bushee et al. (2011) and Green et al. (2014a,
b) examine the determinants and consequences of investor conferences that provide private access to managers. Kirk and
Markov (2016) and Cheng et al. (2016) extend this analysis to analyst/investor days. Subasi (2014) examines institutional investors’
order imbalances during conferences and finds that they buy more stock of future M&A targets than of conference firms that do not
receive bids, suggesting that investors obtain material information about future takeover targets through their interactions at the
conferences. Bushee et al. (2017) finds increases in trade sizes and future trading profits based on trading that occurs in the hours
when managers meet “off-line” with investors at conferences during breakout sessions and one-on-one private meetings.
Solomon and Soltes (2015) examines all of management’s one-on-one meetings with investors for one NYSE firm over a six-year
period (64% of which are at public conference presentations, 21% are roadshows, and 15% are in house meetings). They find that
trades among investors who meet with management in the same quarter better predict future returns than trades executed in quarters
in which meetings did not occur. Using another individual firm, Soltes (2014) examines all private interactions between management
and sell-side analysts. He does not find that private meetings increase the accuracy of analysts’ forecasts and concludes that private
meetings complement other interactions between analysts and managers.
Our corporate jet setting differs from the above papers on conference presentations and analyst days in that jet flights are largely
unobservable to non-participants and do not contain a public disclosure component. Non-participants lack knowledge of the meeting
and are thus unable to modify their trading behavior in anticipation of the meeting (e.g. staying out of the market). By contrast,
market participants excluded from investor conferences can determine whether they are at an information disadvantage by identifying
when the company is presenting and by listening to any public portion of the disclosure. Hence, our setting allows us to identify
unobservable private meetings and examine whether select investors trade on a private meeting that we can identity ex post, but the
broader market could not identify ex ante.
2.2. Interview evidence on corporate jets and private meetings
To better understand the relation between corporate jet flight patterns and private investor meetings, we interviewed a former top
executive of a Fortune 500 company.5 We asked him to describe his experience and understanding of how corporate jets are used for
meetings with investors and sell-side analysts. The executive confirms that CEOs and CFOs often use corporate jets to visit investors
and analysts in money centers. He says that such roadshow trips generally involve flights to multiple money centers over short
windows (e.g., three-to-four days). Thus, focusing on short windows that involve multiple money centers should allow us to isolate
investor and analyst meetings from flights that are taken for operational or personal reasons. Moreover, he mentioned that CEOs and
CFOs often use jets to fly to non-money-center cities that have large concentrations of institutional investors.
The executive gave several motivations for such roadshows, including: (1) inducing institutional investors to purchase more
shares in the company; (2) facilitating the raising of capital; (3) promoting the company’s stock to potential investors; (4) providing
guidance to institutional investors and sell-side analysts while complying with Reg FD; and (5) attending meetings with buy-side
analysts that are organized by sell-side analysts.
These motivations suggest that roadshow flights to meet with investors can be triggered both by new developments in the firm
and by the desire to maintain communications with investors in the absence of news. We will examine whether these motivations are
systematically related to roadshow flights to validate using flight patterns as a proxy for private meetings. However, regardless of the
motivation, roadshow flights allow private meeting participants to update their beliefs, suggesting that significant market reactions
accompany such trips.
4 In 2005, a district court judge dismissed a Reg FD action against Siebel Systems saying (in part) “Regulation FD does not require that corporate officials only utter
verbatim statements that were previously publicly made” (SEC v. Siebel Systems et al. 2005). Subsequent to this decision, only three Reg FD actions have been pursued
by the SEC.
5 This executive’s former company is not in our sample.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
361
2.3. Predictions
The empirical challenge in examining management’s use of corporate jets to meet privately with investors and analysts is that we
cannot directly observe the passengers on jets or the purpose of the trip. Thus, we must infer the purpose and participants from the jet
flight patterns. Our strategy is to use roadshows flight patterns to identify private meetings between managers and analysts/investors.
We focus on trips to the money centers of Boston, Chicago, New York, and San Francisco, as these cities have the highest concentration
of current and potential investors. We also identify flights to other high-ownership cities for the firm. Our interview with
the former executive indicates that flights to meet with investors are often combined on the same extended trip. By contrast, a single
flight to a money center could represent perquisite consumption (e.g., shopping in New York) or an operational purpose (e.g.,
meeting with a customer in Chicago).
We define a roadshow as a three-day flight window with flights to multiple money centers or to multiple high-ownership nonmoney-
center cities. Then, we eliminate any roadshows that are concurrent with a public disclosure (e.g., SEC filing, earnings
announcement, conference presentation) to identify roadshows that are likely used for private meetings with analysts and investors.
As we discuss below, we predict that such roadshow flight windows are significantly associated with stock prices, trading volume,
analyst forecasts, institutional ownership, and potential trading gains if such trips are used for private meetings with investors and
analysts.
First, roadshows provide investors the opportunity to meet with managers to update and revise their beliefs about the firm. Given
concerns about violating Reg FD, managers are unlikely to disclose material quantitative information (e.g., earnings or sales forecasts)
during private meetings with investors. Instead, the stock market reaction to these meetings is more likely to arise from
qualitative information that complements the investors’ private information. Because investors’ ex ante expectations and ex post belief
revisions cannot be measured with precision, we measure the market reaction with abnormal price variability and share turnover
(e.g., Beaver, 1968; Landsman and Maydew, 2002; Cready and Hurtt, 2002). We predict that roadshow flights are associated with
higher abnormal return variability and share turnover than flights to other locations, including flights to a single money center or a
single high-ownership non-money-center city.
Second, our interview suggests that managers often meet with sell-side analysts during roadshows. Similar to investors, analysts
can use the meeting to update and complement their private information about the firm. If such information changes analysts’
expectations about future profitability, we expect that analysts will issue new forecasts soon after the meetings. We measure analyst
forecasting activity using the number of analyst forecasts issued during the flight windows. Because sell-side analysts are more
concentrated in money centers than in the non-money-centers with high institutional ownership, we predict that money center
roadshows are associated with a greater number of new forecasts than non-money-center roadshows and the other flight windows.6
Finding differences in results between money center and non-money-center roadshows with respect to analyst activity would further
validate that our results are driven by private meetings, rather than some more general phenomenon.
Third, roadshows should also affect local institutional ownership in the firm. Bushee et al. (2011) finds sustained increases in
institutional investor ownership subsequent to conference presentations. If roadshows enable investors to meet with managers and
supplement their private information, we predict that roadshow flights to a given metropolitan area are associated with changes in
ownership by institutional investors located in that metropolitan area. Similar to the market reaction tests, it is unclear ex ante
whether the meetings would update investors’ beliefs about the firm positively or negatively. Thus, we examine both signed and
unsigned changes in institutional ownership.
Finally, we provide a direct test of whether the private meetings benefit investors. If private meetings assist investors in correctly
updating their private information about the firm, they should be able to earn trading gains. We estimate local institutional investors’
potential trading gains by multiplying changes in their ownership by the firm’s stock returns in the next quarter. Trading gains are
positive when institutional investors trade in the correct direction and negative when they trade in the opposite direction of future
returns. If roadshows provide beneficial information to local investors, we predict that roadshow flights will be associated with
greater trading gains.
3. Sample and descriptive statistics
3.1. Sample and descriptive statistics for flights
Our sample period is from 2007 to 2010, which are the years that flight data are available on the Wall Street Journal Jet Tracker
database (http://projects.wsj.com/jettracker/).7 For each flight, the database provides the flight operator’s name; the departure date,
time, and location; the arrival date, time, and location; the distance travelled; the tail number; and the approximate cost of the flight.
We construct a list of 5627 unique firms available on Compustat, CRSP, and I/B/E/S between 2007 and 2010 and match each
Compustat firm name on the list to operator names on Jet Tracker. This matching results in 510 jet operator names on Compustat.
6 As mentioned in Section 1, non-money center roadshows are defined based on high institutional ownership, not on high analyst following.
7 Two related studies use the same data source. Yermack (2014) examines the relation between corporate disclosures and CEO vacations. For a sample of 66 CEOs,
he finds that firms tend to disclose good news prior to when CEOs leave on vacation; that stock volatility decreases while the CEO is on vacation; and that stock
volatility increases when the CEO returns from vacation. Lee, Lowry, and Shu (2018) find that flights to subsidiary and plant locations are positively associated with
firm value, which they attribute to enhanced information flow from personal interactions, especially for firms with soft information that is difficult to transmit
remotely.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
362
To identify flights that are most likely used by top management, we require the top arrival location of the jet to be within 100
miles of the firm’s headquarters location as identified in Compustat.8 We also exclude jets operated by subsidiaries. We remove return
flights by eliminating all flights arriving at the top arrival location and/or within 100 miles of the corporate headquarters. We end up
with a total of 395,386 flights for 396 unique firms. On average, each flight covers a distance of 676 miles in 1.6 hours and costs
approximately $4888.9 In dollar terms, the total direct cost of these flights is just slightly below $2 billion.
Panel A of Table 1 reports the total number of flights to the top twenty MSAs ranked in terms of total flights. The New York MSA is
the number one destination, reflecting the large number of businesses and financial institutions in the area, followed by the Chicago,
Washington, and Los Angeles MSAs. There are some “vacation” areas in the top 20, including Miami, Phoenix, and Orlando. About
5% of total flights are to non-US cities, but none of these cities are in the top twenty (London is the most frequent). The second
column shows the number of our sample firms headquartered in each MSA. New York is also the most common headquarters location,
but only accounts for about 5% of the sample firms.
Panel B of Table 1 shows the top twenty locations of institutional investors. We classify institutional investors by MSA using the
ZIP Code listed on their Form 13-F filing. The panel shows that the four money centers are also the top four MSAs in terms of number
of institutional investors (NII) with equity positions in our sample firms during any time in the sample. The last two columns report
the mean number of institutional investors owning the firms’ stock (MNIH) and the mean percentage of institutional holdings (MPIH)
for each MSA during our sample period. These columns show that the average firm in our sample has 119 institutional owners in the
four money center cities that account for about 30% of the holdings in the firms’ stock. The panel shows that other MSAs can have a
large concentration of investors in a given firm’s stock; e.g., Los Angeles and Philadelphia have higher MNIH than San Francisco. We
do not classify these two MSAs as money centers to be consistent with Bushee et al. (2011), but consider such cities in our non-moneycenter
roadshow classification.
We obtain data on firm characteristics from Compustat; stock return and trading volume data from CRSP; analyst following and
forecast data from the I/B/E/S; institutional investor ownership data from the Thomson Reuters Form 13-F database; and debt and
equity issuances from SDC Platinum.
3.2. Classification of flight windows
We use the following algorithm to create non-overlapping flight windows. Our goal is to identify “roadshows” that represent
flights to meet privately with investors; thus, we base the windows around money center flights. We start at the beginning of the data
series for each firm and find the first flight to a money center. Then, we consider that day and the two subsequent trading days as a
three-day money center flight window. Starting with the next date, we search for the next money center flight and create a new threeday
window using the same procedure. We continue through the data series identifying all of the non-overlapping three-day windows
that begin with a money center flight. After identifying all of the money center flight windows, we return to the beginning of the data
series and identify all of the three-day windows that do not contain any flights to money centers, which we define as non-moneycenter
flight windows.
We classify all of the three-day flight windows into five categories. First, we define a money-center roadshow (MC Roadshow) as
any flight window in which there are flights to two or more money centers. Second, we define a single money-center window (MC
Single) as any flight window with flights to only one money center. The MC Roadshow measure gives us more power to detect investor
meetings by separating multiple-city investor roadshows within a short window from one-off trips to money centers (MC Single) that
are more likely taken for personal or operational reasons.
Next, among the non-money-center cities, we define high firm-specific institutional ownership cities as any MSA with at least five
institutional owners or 1% ownership in the firm (these cut-offs are the top quartile of each measure). Then, we identify non-moneycenter
flight windows with flights to two or more firm-specific high-ownership locations and classify them as non-money-center
roadshows (Non-MC Roadshow). Non-money-center flight windows with flights to only one high-ownership city are classified as Non-
MC Single. All remaining flight windows are classified as Other. These windows only contain flights to cities for which the company
has little or no institutional ownership. Thus, these flights are unlikely to contain any investor meetings and are likely for operational
or personal reasons.
Finally, to ensure that our results are not driven by concurrent public disclosures, we exclude flight windows that contain public
news events defined as SEC filings (10-K, 10-Q, 8-K), earnings announcements, and conference presentations. We also search on
Factiva for public notice of presentations and meetings during the 500 roadshow flight windows with the highest three-day Abnormal
ASAR. We find that 11 of them (about 2%) include a notice of a meeting at a location where the corporate jet landed. To obtain a
benchmark of how frequent public disclosures of a meeting occur in our “control” group, we conduct the same search on Factiva for
the 100 other flight windows with the highest Abnormal ASAR. Similar to the roadshow windows, we find that 2% of the flight
windows (2 out of 100) include the announcement or occurrence of a meeting at a location where the corporate jet landed. We
exclude these 13 observations from our sample.
We illustrate the flight windows with an example for the 3M Company in Appendix A. We classify flight windows containing
8 There are 33 pairs for which the top arrival location of the operator’s jets is not within 100 miles of the matched firm’s headquarters location (for non-U.S.
companies, we use the location of the U.S. corporate headquarters from the company’s website). We include these in the sample because the name matches are almost
exact and we cannot find any evidence indicating that the operator and the firm are two different companies.
9 The reported values are the means. The cost is based on Wall Street Journal’s estimated cost to operate each flight. The detailed estimation process is available on
the Wall Street Journal Jet Tracker website.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
363
money center cities as either MC Roadshow (e.g., the 10/03/2007 window with flights to Boston and Chicago) or MC Single (e.g., the
10/08/2007 window with flights only to Chicago). The “Non-MC Cities” column has all of the flights to non-money-center MSAs with
high institutional investor ownership in 3M (e.g., at least five institutional owners or 1% ownership in the firm). The 11/19/2007
window has flights to two such cities (Miami and Milwaukee) and no flights to money centers; thus, we classify it as a Non-MC
Roadshow. The 10/31/2007 window has a flight to one high-ownership city (Los Angeles) and no flights to money centers; thus, we
classify it as a Non-MC Single. All remaining windows have only flights to cities with little or no institutional ownership in 3M and are
classified as Other.
This example shows that the money-center flight windows often have flights to other high-ownership non-money-center cities and
other cities. However, if investors meetings explain the results, we expect our empirical results to be driven by the money center
flights. Also, money-center roadshows can also contain non-money roadshows (see the 11/12/2007 window). We classify these as MC
Roadshow because the money centers have greater numbers of investors for the firm in almost every case. Also, this decision ensures
Table 1
Flight and investor locales.
Panel A: Number of flights to top-20 destination MSAs
Rank Location Number of flights Number of firms HQ in MSA
1 New York-Northern New Jersey-Long Island, NY-NJ-PA MSA 32,068 22
2 Chicago-Joliet-Naperville, IL-IN-WI MSA 17,382 12
3 Washington-Arlington-Alexandria, DC-VA-MD-WV MSA 16,221 5
4 Los Angeles-Long Beach-Santa Ana, CA MSA 11,197 8
5 Atlanta-Sandy Springs-Marietta, GA MSA 9784 13
6 Dallas-Fort Worth-Arlington, TX MSA 9404 12
7 Houston-Sugar Land-Baytown, TX MSA 8057 16
8 Miami-Fort Lauderdale-Pompano Beach, FL MSA 7675 3
9 Boston-Cambridge-Quincy, MA-NH MSA 7504 6
10 Philadelphia-Camden-Wilmington, PA-NJ-DE-MD MSA 6390 8
11 Minneapolis-St. Paul-Bloomington, MN-WI MSA 5098 9
12 Baltimore-Towson, MD MSA 4595 2
13 Phoenix-Mesa-Glendale, AZ MSA 4588 2
14 San Francisco-Oakland-Fremont, CA MSA 4458 11
15 Denver-Aurora-Broomfield, CO MSA 4456 8
16 St. Louis, MO-IL MSA 3774 7
17 Pittsburgh, PA MSA 3705 5
18 Detroit-Warren-Livonia, MI MSA 3573 4
19 Orlando-Kissimmee-Sanford, FL MSA 3535 1
20 Columbus, OH MSA 3179 5
Panel B: Geographic location of institutional investors
Rank Location NII MNIH MPIH
1 New York-Northern New Jersey-Long Island, NY-NJ-PA MSA 795 59.748 0.128
2 Boston-Cambridge-Quincy, MA-NH MSA 202 26.003 0.108
3 Chicago-Joliet-Naperville, IL-IN-WI MSA 170 20.424 0.039
4 San Francisco-Oakland-Fremont, CA MSA 164 12.489 0.023
Total for top 4 locations 1331 118.664 0.298
5 Los Angeles-Long Beach-Santa Ana, CA MSA 141 12.569 0.055
6 Bridgeport-Stamford-Norwalk, CT MSA 124 8.621 0.010
7 Philadelphia-Camden-Wilmington, PA-NJ-DE-MD MSA 109 12.954 0.012
8 Dallas-Fort Worth-Arlington, TX MSA 67 6.636 0.008
9 Washington-Arlington-Alexandria, DC-VA-MD-WV MSA 61 7.112 0.004
10 Minneapolis-St. Paul-Bloomington, MN-WI MSA 48 5.823 0.005
11 Houston-Sugar Land-Baytown, TX MSA 47 4.661 0.002
12 Atlanta-Sandy Springs-Marietta, GA MSA 36 5.133 0.012
13 Baltimore-Towson, MD MSA 33 4.633 0.022
14 San Diego-Carlsbad-San Marcos, CA MSA 32 3.795 0.005
15 Seattle-Tacoma-Bellevue, WA MSA 32 3.227 0.004
16 Cincinnati-Middletown, OH-KY-IN MSA 31 5.451 0.002
17 Milwaukee-Waukesha-West Allis, WI MSA 31 5.127 0.006
18 Denver-Aurora-Broomfield, CO MSA 29 5.793 0.014
19 St. Louis, MO-IL MSA 29 4.326 0.002
20 Detroit-Warren-Livonia, MI MSA 25 3.954 0.002
This table presents the top destinations of corporate jet flights along with the geographic locations of firms’ headquarters (HQ) and institutional investors. We base
flight destinations and investor locations on Metropolitan Statistical Areas (MSAs). Panel A lists the top 20 destinations for flights in our sample firms and the number
of firms headquartered in each MSA. Panel B presents the geographic locations of institutional investors that own shares in our sample firms. NII is the number of
unique institutional investors with positive shareholdings in at least one of our sample firms. MNIH is the mean number of institutions in the MSA that hold shares in
our sample firms. MPIH is the mean percentage shares outstanding held by institutions located in the MSA.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
364
that the Non-MC Roadshow classification does not include any money center flights, which provides a strong validity check that
investor meetings, rather than some other phenomenon associated with money center flights, explain the results.10
3.3. Validation of flight windows as a proxy for investor meetings
To validate roadshow flights as a proxy for private meetings with investors, we examine the relation between roadshows and
incentives for managers to fly to privately meet investors.11 Managers likely prefer private disclosure when their information may be
misinterpreted by unsophisticated investors; when investors demand privileged access; and when there are changes in the firm’s
information environment (Tasker, 1998; Bushee et al., 2003). We include the ratio of intangible assets to total assets (Intangible) to
proxy for the complexity of the firm’s information. We use the percentage of institutional investor holdings (Institutional Ownership) to
proxy for investor demands for privileged access to management.
We include five variables to capture changes in a firm’s information environment that would likely be associated with institutional
investor demands to meet with managers. We use analyst forecast dispersion (Analyst Forecast Dispersion) and the change in short
interest (△Short Interest) to broadly capture changes in a firm’s information environment. For example, institutions may demand more
information (and want to meet with management) when short interest increases to discern whether short sellers have private information
about the firm. Changes in a firm’s information environment can also be triggered by events such as earnings announcements
and security issuances. For example, managerial incentives to engage in disclosure and investor relations activities are
likely stronger prior to equity and debt issuances (Lang and Lundholm, 2000; Shroff et al. 2013). Therefore, we include an indicator
variable for any debt or equity issue that is in the subsequent month (Debt/Equity Issue). We also include an indicator variable for
whether there was an earnings announcement in the prior month (Prior Earn Announcement) and the market-adjusted return around
the most recent earnings announcement (Announcement Ret) to proxy for potential changes in institutional demand to meet with
managers around these events.
As discussed by Bushee and Miller (2012), managers sometimes commit to keep select investors informed about the firm.
Therefore, managers who commit to a number of costly flights to meet directly with key investors, regardless of news, signal a greater
commitment to transparency to those investors. We measure the manager’s commitment to keep selective investors informed by using
the log of the number of roadshows in the same month in the prior year (Log Prior MC Roadshow and Log Prior Non-MC Roadshow).
We also include a number of firm characteristics that relate to a firm’s information environment or general motivations for flight
activity. We do not predict the signs of the control variables because their effect on the firm’s money center flight activity is unclear.
These controls include firm size (Log MVE); book-to-market ratio (BM Ratio); sales growth (Sales Growth); the change in net income
(△Earn); earnings per share scaled by price (EP Ratio); analyst following (Log Analyst Coverage); the leverage ratio (Leverage); beta
(Beta); the standard deviation of stock returns (Std Dev); prior market-adjusted stock returns (Returns); prior share turnover (Share
Turnover); the total number of flights in the same month (Log Total Flight); and prior year flight activity (Log Prior Total Flight).
Finally, we include proxies for public disclosure events, such as the log of the number of money center (Log MC Conference) and
non-money-center (Log Non-MC Conference) conference presentations as well as the log of the number of Form 8-K filings during the
month (Log 8K). Because a greater number of public disclosures should “crowd out” the available time for private meeting flights, we
expect a negative relation between these public events and our private roadshow flight variables.
Table 2 presents descriptive statistics for the variables. We winsorize all continuous variables at the 1st and 99th percentiles. The
sample firms tend to be large with high institutional ownership (median=75%) and analyst following (median=5), suggesting that
larger firms are more likely to own a fleet of corporate jets.
Table 3 presents results from regressions of the roadshow windows on proxies for motivations to meet with investors. This
analysis is conducted at the firm-month level, resulting in a sample size of 9510. In columns (1) and (2), the dependent variables are
the log of the number of MC Roadshow windows for the firm in a month (Log MC Roadshow) and the log of the number of Non-MC
Roadshow windows (Log Non-MC Roadshow). In column (3), the dependent variable is log of the number of flights to other locations
(Log Other Flight), which excludes flights to money centers or firm-specific high-ownership locations. If our roadshow variables
capture private meetings with investors rather than general flight motivations, then the signs and/or significance levels on the
determinant variables measuring motivations to meet with investors should differ in this regression.
Column (1) shows that firms with higher intangible assets and higher institutional ownership have a significantly greater number
of money center roadshows, indicating that managers are more likely to fly to meet with investors when the firm’s financial information
environment is more complex and investor demands for privileged access are high. There are also significantly more
money center roadshows in the month immediately after an earnings announcement and for firms about to raise capital. These results
suggest that managers are also more likely to meet with institutional investors when there are potential changes in the firm’s information
environment arising from firm-specific events. Finally, MC Roadshow increases in the number of money center roadshows
in the same month of the prior year, indicating that managers use roadshows to commit to a higher level of transparency to selected
key investors.
10 Among the MC Roadshow windows, 42% have no flights to other high-ownership cities, 27% have flights to one other high-ownership city, and 31% have flights
to multiple other high-ownership cities. Thus, money center roadshows likely often include visits to investors in other cities as well. Our decision to group such
roadshows as only MC Roadshow likely weakens the results for Non-MC Roadshow windows. Thus, any significant results we do find for Non-MC Roadshow windows
increases the likelihood that we identify private meetings.
11 We do not assume that jet trips are a complete substitute for phone calls; managers likely use both. Rather, we are discussing incentives for the managers to not
solely rely on low-cost distance communications.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
365
In column (2), we replace MC Roadshow with Non-MC Roadshow as the dependent variable. Our results indicate that non-money
center roadshows are positively associated with institutional ownership, upcoming security issuances, and the number of non-money
center roadshows in the same month of the prior year. In contrast with money center roadshows, non-money center roadshows are
not significantly associated with the level of intangibles, and are lower in the month immediately after an earnings announcement.
Thus, non-money center road shows appear to be primarily motivated by a commitment to transparency with the institutional
investors, rather than by responding to the complexity of information or recent news. We also note that the number of conference
presentations and Form 8-K filings are negatively associated with both money center and non-money center roadshows, which
suggests that public information events crowd out opportunities for private meeting roadshows in a month, further validating that
our roadshows are picking up purely private meetings.12
In contrast with columns (1) and (2), column (3) shows a negative significant relation between other flights and intangible assets,
percentage of institutional ownership, and prior money center roadshow flights. Moreover, there is no significant relation between
other flights and debt/equity issuance or earnings announcement month. Thus, the determinants of other flights differ from the
determinants of roadshows. Overall, the results suggest that the roadshow variables identify private meetings between managers and
investors.
Table 2
Descriptive statistics.
Variable N Mean Std Dev Median P25 P75
No. MC Roadshow 9510 0.385 0.770 0.000 0.000 1.000
Log MC Roadshow 9510 0.225 0.406 0.000 0.000 0.693
No. Non-MC Roadshow 9510 0.278 0.649 0.000 0.000 0.000
Log Non-MC Roadshow 9510 0.159 0.367 0.000 0.000 0.000
No. Other Flight 9510 13.729 15.523 9.000 4.000 18.000
Log Other Flight 9510 2.183 1.091 2.303 1.609 2.944
Intangible 9510 0.179 0.185 0.112 0.020 0.298
Institutional Ownership 9510 0.715 0.182 0.745 0.641 0.831
Analyst Forecast Dispersion 9510 0.047 0.267 0.028 0.011 0.070
Δ Short Interest 9510 0.000 0.010 0.000 −0.004 0.004
Debt/Equity Issue 9510 0.069 0.253 0.000 0.000 0.000
Prior Earn Announcement 9510 0.349 0.477 0.000 0.000 1.000
Announcement Ret 9510 0.001 0.058 0.001 −0.026 0.028
MVE 9510 19,397 36,420 5196 1717 18,630
Log MVE 9510 8.614 1.676 8.556 7.448 9.833
BM Ratio 9510 0.647 0.466 0.539 0.341 0.821
Sales Growth 9510 0.020 0.229 0.023 −0.090 0.116
Δ Earn 9510 0.000 0.028 0.000 −0.006 0.005
EP Ratio 9510 −0.001 0.084 0.015 0.007 0.022
Leverage 9510 0.249 0.171 0.234 0.119 0.358
Beta 9510 1.180 0.594 1.079 0.761 1.496
Std Dev 9510 0.032 0.016 0.028 0.019 0.041
Analyst Coverage 9510 6.899 6.283 5.000 2.000 10.000
Log Analyst Coverage 9510 1.717 0.893 1.792 1.099 2.398
Returns 9510 0.047 0.374 0.006 −0.178 0.197
Share Turnover 9510 0.273 0.178 0.227 0.158 0.336
No. 8k 9510 3.050 2.377 2.000 2.000 4.000
Log 8K 9510 0.911 0.749 1.099 0.000 1.386
No. MC conference 9510 0.295 0.624 0.000 0.000 0.000
Log MC conference 9510 0.134 0.318 0.000 0.000 0.000
No. Non-MC Conference 9510 0.161 0.440 0.000 0.000 0.000
Log Non-MC Conference 9510 0.076 0.239 0.000 0.000 0.000
This table presents descriptive statistics for the variables used in our tests conducted at the firm-month level that validate the empirical proxies for private meetings.
Variable definitions are provided in Appendix B. All continuous variables are winsorized at the 1st and 99th percentiles.
12 We also partition roadshows into routine and non-routine roadshows based on historical flight patterns. However, our data span a fairly short time period
(2007–2010) making it challenging to classify roadshows as routine with minimal measurement error. Nevertheless, in untabulated analysis, we define a roadshow as
non-routine if roadshows are not present during the same week and month in the other three years. Our results indicate that Log Prior Roadshow and Institutional
Ownership are positively associated with both routine and non-routine roadshows. Debt/Equity Issue is positively associated with routine roadshows but not associated
with non-routine roadshows. However, there is no significant difference between the coefficients in the two regressions. Both Intangible and Prior Earn Announcement
are not associated with either routine or non-routine roadshows, potentially because this classification does not separate out money center and non-money center
roadshows.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
366
Table 3
Determinants of roadshow and other flights.
Log MC roadshow Log Non-MC roadshow Log other flights
Intangible 0.112** −0.075 −0.550**
(2.426) (−1.308) (−2.426)
Institutional Ownership 0.099*** 0.087*** −0.356***
(4.621) (4.501) (−7.144)
Analyst Forecast Dispersion −0.003 −0.013 −0.014
(−0.216) (−1.008) (−0.257)
Δ Short Interest −0.202 −0.049 −0.053
(−0.628) (−0.229) (−0.160)
Debt/Equity Issue 0.028** 0.027** −0.024
(2.127) (2.220) (−0.785)
Prior Earn Announcement 0.020** −0.024** −0.063
(1.980) (−1.976) (−1.536)
Announcement Ret −0.013 0.060 0.299
(−0.199) (0.831) (1.364)
Log Prior MC Roadshow 0.245*** −0.454***
(10.873) (−6.335)
Log Prior Non-MC Roadshow 0.250*** 0.082
(7.104) (1.097)
Log MVE 0.018*** 0.051*** −0.190***
(2.963) (5.964) (−5.876)
BM Ratio 0.003 −0.002 −0.038
(0.175) (−0.085) (−0.575)
Sales Growth −0.019 0.016 0.044
(−0.901) (0.673) (0.520)
Δ Earn 0.233 0.209 0.300
(1.490) (1.606) (0.833)
EP Ratio −0.021 −0.075 −0.136
(−0.485) (−1.258) (−0.741)
Leverage −0.013 −0.076 −0.049
(−0.294) (−1.215) (−0.208)
Beta 0.000 −0.023 −0.046
0.027 (−1.346) (−0.804)
Std Dev 0.329 1.019⁎ −2.005
(0.660) (1.901) (−1.132)
Log Analyst Coverage −0.007 0.011 0.024
(−1.042) (1.041) (0.617)
Returns −0.012 −0.023⁎ 0.106⁎
(−0.933) (−1.873) (1.933)
Share Turnover −0.023 −0.127*** 0.038
(−0.622) (−2.712) (0.224)
Log 8K −0.039*** −0.014*** −0.047**
(−6.685) (−2.983) (−2.560)
Log MC conference −0.042*** −0.037*** −0.125***
(−3.144) (−2.622) (−3.596)
Log Non-MC Conference −0.063*** −0.035** 0.013
(−3.601) (−2.077) (0.223)
Log Total Flight 0.169*** 0.082*** 0.893***
(14.844) (7.472) (24.343)
Log Prior Total Flight −0.008 −0.001 0.024
(−0.757) (−0.128) (0.734)
Intercept −0.496*** −0.539*** 1.703***
(−6.739) (−5.570) (5.194)
N 9510 9510 9510
Adj R2 0.271 0.251 0.470
*,**,*** Significantly different from zero at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
This table presents tests that validate money center roadshows and non-money-center high institutional ownership roadshows as empirical proxies of private meetings.
The regressions are estimated at the monthly firm level. In the first column, the dependent variable is Log MC Roadshow, which is measured as log(1+number of
money center roadshows during the month). In the second column, the dependent variable is Log Non-MC Roadshow, which is measured as log(1+number of nonmoney-
center roadshows during the month). In the third column, the dependent variable is Log Other Flights, which is measured as log(1+number of flights during the
month to non-money center cities that are not high firm-specific institutional ownership cities). The regressions include the following determinants of monthly flights.
Intangible is the ratio of the firm’s intangible assets to total assets at the end of the prior fiscal quarter. Institutional Ownership is the percentage of shares outstanding
held by institutional investors at the end of the prior calendar quarter. Analyst Forecast Dispersion is the standard deviation of analyst EPS forecast scaled by the mean
analyst EPS forecast in the prior month. Δ Short Interest is the month-to-month change in the ratio of shares sold short to trading volume measured in the prior month.
Debt/Equity Issue is an indicator coded as 1 if there is either a debt or equity issuance in the subsequent month and 0 otherwise. Prior Earn Announcement is an indicator
variable equal to 1 if the firm announces earnings in the prior month. Announcement Ret is the value-weighted market-adjusted return for the three-day window
(−1,+1) centered around the earnings announcement date. Variable definitions are provided in Appendix B. Standard errors are clustered at both the firm- and
month-level. All continuous variables are winsorized at the 1st and 99th percentiles.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
367
4. Market reactions to flights
4.1. Stock market reaction during flight windows
We examine stock market reactions to corporate jet flights using abnormal absolute returns and abnormal share turnover in threeday
flight windows. We compute the standardized absolute value of size-adjusted returns (Abnormal ASAR) as the difference between
three-day absolute size-adjusted returns and the mean three-day absolute size-adjusted returns in an estimation period, divided by the
standard deviation of mean absolute size-adjusted returns in the estimation period (Subramaniam, 1995; Cready and Hurtt, 2002).13
We measure abnormal share turnover (Abnormal Turnover) as three-day volume divided by shares outstanding, less the average threeday
turnover in the estimation period, multiplied by 100.14
Table 4 presents results from regressions of Abnormal ASAR and Abnormal Turnover on the flight window indicators representing moneycenter
roadshows (MC Roadshow), windows with flights to one money-center (MC Single), high-ownership non-money-center roadshows (Non-
MC Roadshow), and windows with flights to one high-ownership non-money center location (Non-MC Single). Flight windows with flights to
other cities for which a firm has little or no institutional ownership are the omitted category. The regressions include the control variables from
Table 3.15 After requiring data for all of the variables, we have a sample of 44,413 three-day flight windows. For all regressions, t-statistics are
in parentheses and are calculated using standard errors clustered at both the firm- and quarter-level.16
The first two columns show that the coefficient on the MC Roadshow indicator is positive and significant in both the Abnormal
ASAR and Abnormal Turnover regressions. Moreover, the coefficients on MC Roadshow are significantly greater than the coefficients
on the indicator for flights to a single money center (MC Single), which is only significant in the turnover regression. These results
suggest that the largest return and volume reactions accompany roadshows that involve flights to multiple money centers in a threeday
window, consistent with the roadshow measure being a proxy for private meetings that generate significant price and volume
reactions.17 However, the economic magnitudes are small. The coefficient on Abnormal ASAR (Abnormal Turnover) of 0.050 (0.134)
represents a 4.0% (3.6%) relative increase in absolute ASAR (share turnover) for money-center roadshows compared to other flight
windows.18 The small economic magnitudes are not surprising given that only a small subset of investors meet with managers; other
investors are unaware of the meeting and could be trading to “correct” what appears to be an unwarranted price movement.
Moreover, the manager is unlikely to be releasing material new public information during these meetings; rather, the market reaction
reflects a small subset of investors updating their private information.
The results also show that high-ownership non-money-center roadshows (Non-MC Roadshow) are associated with greater return
volatility and share turnover than single flights (Non-MC Single) to high-ownership cities and to cities with little or no institutional
ownership in the firm. These results again suggest that flights to multiple non-money-center cities with high firm-specific ownership
capture unobservable private meetings with large investors in multiple locations.19
One possible explanation for these results is that cities with high firm-specific institutional ownership proxy for other large cities
where managers fly for other reasons. We address this possibility by including indicators for multiple (Multiple Large Cities) or single
(Large City) flights to any of the next ten MSAs with the most frequent number of flights.20 The results show that flights to the next ten
largest cities are not significantly associated with either Abnormal ASAR or Abnormal Turnover.
We next examine the signed stock returns during the flight windows. One possible explanation for the above results is that managers
use the corporate jet to meet with investors when the firm is experiencing an uptick in return volatility or volume for whatever reason. In
this scenario, we would expect to see more roadshows in response to negative news than to positive news. If a manager observed negative
returns that he perceived to be unjustified, he would have incentives to meet with key investors to correct the market’s assessment of firm
value. However, positive returns would not provide a clear motivation for managers to immediately use the jet for investor meetings. In
13 We measure size-adjusted returns as the difference between firm i’s returns and the size-decile-matched CRSP portfolio returns over the same event and estimation
periods. The estimation period spans two 45-day periods before and after the flight window [(-60,-16) and (+16,+60)]. We also used an estimation period 90-
days prior to the flight window (-105,-16) but found that, over the 2007 to 2010 period, volatility and volume steadily trended up, which biased toward finding
abnormal volatility and volume.
14 We are unable to pinpoint the exact hour during which the private meetings take place. Thus, we are unable to use TAQ data to examine intraday trade sizes or
trading profits. However, such precise timing is less of a concern in our study because we eliminated any flight windows with public disclosures. Thus, the entire
trading day represents a period when private meetings potentially occur in the absence of public disclosures.
15 We make some modifications to tailor the variables to the research design for these tests. We drop the public event variables (conference presentations and Form
8-K) since they are zero for all flight window and we measure prior number of flights using the past 21 days.
16 We also estimated all of the regressions with (1) firm fixed effects, clustering standard errors by quarter and (2) quarter fixed effects, clustering by firm. All of our
coefficients of interest retain their statistical significance at comparable levels under these alternative approaches.
17 To alleviate concerns that the results could be driven by differences in unobserved firm characteristics between quarters in which firms only flew to either money
centers or non-money-centers, we restrict the sample to firms that have both money center and non-money-center flights in the same quarter. The results are
quantitatively similar.
18 While our standardized measure of Abnormal ASAR has desirable statistical properties (Cready and Hurtt 2002), the magnitude is difficult to interpret. Multiplying
the ABS_SAR by 2.0% (the mean standard deviation in the estimation period) yields the difference between the mean absolute SAR in the event window and the
mean in the estimation window, which is 2.5%. Thus, a mean ABS_SAR of 0.050 represents an incremental absolute SAR of 0.10% (10 basis points), which is a 4.0%
increase over the estimation period value (0.10% / 2.5%).
19 We also dropped the “money center” designation and classified all flight windows by the amount of firm-specific local institutional ownership. Flights and
roadshows to these top firm-specific ownership cities under this modified definition continue to significantly explain return volatility and share turnover.
20 As shown in Table 1, these MSAs are Washington, Los Angeles, Atlanta, Dallas, Houston, Miami, Philadelphia, Minneapolis-St. Paul, Baltimore, and Phoenix. We
found the same results with indicators for the next top five cities.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
368
this case, our explanation that investor meetings trigger the stock return reaction is more plausible.
In column (3), we estimate OLS regression with signed size-adjusted returns as the dependent variable. The coefficient on MC
Roadshow is positive and significant, indicating that money center roadshows are associated with positive news, on average; although
the magnitude of the reaction is small (0.2%). The coefficient on Non-MC Roadshow is also positive and significant, with the magnitude
around 0.1%. These small magnitudes are not surprising given that managers are unlikely to disclose material information
privately after Reg FD; rather, the price reaction likely reflects the meetings allowing investors to update their private information.
These findings suggest that managers are more likely to undertake roadshows when they believe the stock price is undervalued,
which is more consistent with the investor meetings triggering the price reaction than an alternative story that a day 0 price
movement leads managers to fly to investors to manage the news.
The positive signed return reaction to the roadshow flight windows could reflect either a permanent price change due to impounding new
information or a transitory price change due to hype or overreaction caused by the meetings. To provide evidence on these possibilities, we
Table 4
Stock market reactions during three-day flight windows.
Abnormal ASAR Abnormal turnover SAR
MC Single 0.022 0.068*** 0.000
(1.300) (2.717) (0.439)
MC Roadshow 0.050***†† 0.134***††† 0.002**††
(3.521) (3.558) (2.253)
Non-MC Single 0.007 −0.005 0.000
(0.511) (−0.246) (−0.498)
Non-MC Roadshow 0.037**† 0.070**††† 0.001**†
(2.019) (2.130) (2.153)
Large City −0.012 0.001 −0.001
(−0.626) (0.021) (−0.578)
Multiple Large Cities 0.010 0.061 −0.001
(0.204) (0.945) (−0.361)
Log Prior MC Flight −0.006 −0.045*** 0.000⁎
(−0.798) (−2.792) (−1.698)
Log Prior Non-MC Flight −0.008 −0.002 0.000
(−1.070) (−0.141) (0.129)
Institutional Ownership −0.012 −0.007 −0.003***
(−0.709) (−0.072) (−3.047)
Log Analyst Coverage 0.000 0.011 0.000
(0.013) (0.532) (0.134)
Log MVE 0.002 0.010 0.000
(0.474) (0.724) (−0.291)
EP Ratio −0.029 0.860 0.024
(−0.228) (1.096) (1.559)
BM Ratio 0.005 −0.033 0.002***
(0.376) (−1.044) (3.192)
Leverage −0.009 −0.100⁎ 0.000
(−0.473) (−1.933) (−0.096)
Intangible −0.035** −0.050 0.000
(−2.469) (−0.996) (−0.509)
Returns −0.013 0.073⁎ −0.001
(−1.023) (1.903) (−0.642)
Share Turnover −0.001 −0.953*** 0.000
(−0.037) (−5.284) (−0.192)
Δ Earn 0.024 −0.606 0.000
(0.177) (−0.898) (−0.018)
Sales Growth 0.001 0.037 0.000
(0.068) (0.456) (0.044)
Beta −0.007 −0.018 0.000
(−0.564) (−0.349) (0.407)
Std Dev −0.322 1.230 0.069
(−0.372) (0.487) (1.419)
Prior Earn Announcement −0.044 −0.059 0.000
(−1.517) (−0.703) (0.203)
Analyst Forecast Dispersion −0.003 0.121⁎ 0.003
(−0.211) (1.656) (1.399)
Δ Short Interest −0.860 −0.350 0.038
(−0.789) (−0.159) (0.778)
Debt/Equity Issue 0.005 0.109** 0.000
(0.311) (2.322) (−0.025)
Announcement Ret −0.151 −0.328 −0.003
(−1.579) (−1.103) (−0.367)
Intercept −0.027 0.011 0.000
(continued on next page)
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
369
examine return windows subsequent to the flight windows in untabluated tests. We re-estimate the regression in Table 4 column (3) using sizeadjusted
returns starting three days after the start of the flight window and extending over various horizons up to 60 days later. We find no
evidence of significant return reversals over any horizon. The three-day signed return reactions during the flight windows do not, therefore,
appear to be due to overreaction or hype, but rather reflect a permanent price movement.21
Finally, in untabulated tests, we repeat the analyses using rank regressions to ensure that skewness does not drive our results. The
distributions of Abnormal ASAR and Abnormal Turnover are skewed—untabulated univariate evidence shows positive means and
negative medians for both variables during roadshow windows, with only about 40% of roadshows experiencing increases in absolute
returns or volume. Our inferences remain the same when we use rank regressions.22
4.2. Analyst coverage during flight windows
While our focus is on private meetings with investors, managers also likely take the opportunity to meet with sell-side analysts
during these roadshows. There is anecdotal evidence that “management travels to the offices of buy-side and sometimes sell-side
analysts to discuss the financial and strategic prospects for the company” (Ryan and Jacobs, 2008, Kindle Locations 725–729). We
expect meetings with sell-side analysts to be more likely to occur in the case of MC Roadshow because of the high concentration of
analysts in money centers. By contrast, we do not expect Non-MC Roadshow to be a reasonable proxy for private meetings with
analysts. Thus, finding differences in results between MC Roadshow and Non-MC Roadshow would further validate that our results are
driven by private meetings, rather than by some more general phenomenon.
Table 5 presents the results of this analysis. We measure analyst forecasting activity (Log Analyst Forecast) using the number of
analyst forecasts issued during the flight windows (0,+2). We control for the number of analyst forecasts prior to the flights (−90,
−1), as well as the other control variables from Table 5.23 As expected, the number of analysts who issued forecasts during the flight
window is positively associated with MC Roadshow, but not associated with Non-MC Roadshow. Notably, the coefficient on MC Single
is positive and significant, indicating that flights to only one money center likely contain some meetings with analysts, which
encourage them to issue new forecasts. These results, together with the results in Table 4, suggest that market reactions to roadshows
are likely driven by private meetings rather than by some more general phenomenon.
Table 4 (continued)
Abnormal ASAR Abnormal turnover SAR
(−0.531) (0.085) (−0.095)
N 44,413 44,413 44,413
Adj R2 0.010 0.030 0.010
*,**,*** Significantly different from zero at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
†, ††, ††† MC Roadshow (Non-MC Roadshow) significantly greater than MC Single (Non-MC Single) at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
This table presents regressions that test the relation between flights and stock market reactions. The unit of observation is the three-day flight window (see
Appendix A). In column (1), the dependent variable is abnormal absolute value size-adjusted returns (Abnormal ASAR), which is measured as the difference between
the three-day absolute size-adjusted returns and the mean three-day absolute size-adjusted returns over a 90-day estimation period divided by the standard deviation
of the mean absolute size-adjusted returns over the estimation period. In column (2), the dependent variable is abnormal share turnover (Abnormal Turnover), which is
measured as the three-day volume divided by shares outstanding less the average three-day turnover in the 90-day estimation period, multi abnormal share turnover is
measured as the three-day volume divided by shares outstanding less the average three-day turnover in the 90-day estimation period, multiplied by 100. In column (3),
the dependent variable is the three-day size-adjusted return during the flight window (SAR). We identify three-day flight windows using indicator variables. MC Single
is an indicator variable coded as 1 for three-day window with flights to only one money center, and 0 otherwise. MC Roadshow is an indicator variable coded as 1 for
three-day window with flights to two or more money centers, and 0 otherwise. We define money centers as locations in the Boston, Chicago, New York, and San
Francisco MSAs. See Appendix A for more details. Non-MC Single is an indicator variable coded as 1 for three-day window with flights to a single city where the firm
has a high level of firm-specific institutional ownership, and 0 otherwise. Non-MC Roadshow is an indicator variable coded as 1 for three-day window with flights to
two or more cities where the firm has a high level of firm-specific institutional ownership, and 0 otherwise. We define high firm-specific institutional ownership cities
as any MSA with at least five institutional owners or 1% ownership in the firm. Large City is an indicator variable coded as 1 if there is one flight to any of the next ten
MSAs with the most frequent number of flights in the three-day window, and 0 otherwise. Multiple Large Cities is an indicator variable coded as 1 if there are flights to
two or more cities in the next ten MSAs with the most frequent number of flights in the three-day window, and 0 otherwise. The omitted category is three-day window
with flights to other cities for which a firm has low or no institutional ownership. Variable definitions are provided in Appendix B with the following modifications to
tailor the variables more closely to the research design for these tests. To control for the possibility that the market reaction in the three-day window is a delayed (or
pre-empted) reaction from earlier flights, Log Prior MC Flight is measured as the log of one plus the number of money center flights in the prior 21 days and Log Prior
Non-MC Flight is measured as the log of one plus the number of non-money-center flights in the prior 21 days. Standard errors clustered at both the firm- and quarterlevel.
All continuous variables are winsorized at the 1st and 99th percentiles.
21 Another possible alternative explanation is that the results reflect delayed or pre-emptive trading for public announcements that occur on the day before or the
day after the three-day flight window. We eliminated all flight windows with a public event on the day before or day after the flight window, which reduced the sample
by 6,266 observations. The results for the Table 4 columns (1) and (2) regressions are identical. For the column (3) regression, the coefficient on Non-MC Roadshow is
no longer significant. Thus, adjacent public event days are unlikely to explain our results.
22 As another alternative specification, we adjust returns using the Fama and French daily four factors (MKT, SMB, HML and UMD) and exclude controls for firm
characteristics such as size, beta, and book-to-market. Our results are robust to this alternative specification.
23 Because we control for the number of analyst forecasts in the 90-day period prior to the flight window, we do not control for analyst following (Log Analyst
Coverage) but note that the results remain unchanged when we include Log Analyst Coverage as an additional control variable.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
370
5. Changes in institutional ownership and trading gains
5.1. Changes in institutional ownership
In this section, we examine whether roadshow flights to a given MSA are associated with quarterly changes in ownership by
institutional investors located in that MSA. As it is unclear which direction the meetings update investors’ beliefs about the firm, we
examine both signed and unsigned changes in institutional ownership. For each money center, we measure the change in number of
roadshows (Δ Log MC Roadshow) as the difference in the log of the number of money center roadshows in this quarter and the prior
quarter, and the change in single money center flights (Δ Log MC Single) as the difference in the log of the number of single money
center flight windows. We apply the same procedure for non-money center high institutional ownership locations (Δ Log Non-MC
Roadshow and Δ Log Non-MC Single) and large cities that are not high-ownership locations for the firm (Δ Log Large City and Δ Log
Multiple Large Cities). This analysis is conducted at the firm-quarter-MSA level, yielding a sample size of 291,225.
In Table 6, column (1) presents results for signed changes in local institutional ownership as the dependent variable while column
(2) presents results for unsigned changes. We include quarter fixed effects and the same control variables that are used in Table 4, and
cluster standard errors at the MSA and firm-quarter levels. In column (1), the coefficient on Δ Log MC Roadshow is positive and
significant, implying that more money center roadshows are associated with larger increases in ownership by local institutions. Not
surprisingly, the coefficient on Δ Log MC Roadshow is also positive and significant in column (2). In terms of economic significance, an
increase from zero to one money center roadshow over the quarter is associated with a 3.6% increase in local institutional ownership.
Table 5
Analyst forecasting activity during three-day flight windows.
Log analyst forecast
MC Single 0.041***
(3.853)
MC Roadshow 0.030**
(2.533)
Non-MC Single 0.009
(0.925)
Non-MC Roadshow 0.006
(0.346)
Large City 0.002
(0.190)
Multiple Large Cities −0.022
(−1.022)
Log Prior MC Flight −0.040***
(−3.460)
Log Prior Non-MC Flight 0.026**
(2.267)
Institutional Ownership 0.061
(1.313)
Log Prior Analyst Forecast 0.147***
(9.299)
Log MVE 0.058***
(6.545)
EP Ratio 0.139**
(2.248)
BM Ratio 0.043**
(2.215)
Leverage −0.020
(−0.434)
Intangible −0.237***
(−5.119)
Returns −0.017
(−1.418)
Share Turnover 0.084
(1.372)
Δ Earn −0.422***
(−3.010)
Sales Growth 0.069***
(3.125)
Beta −0.009
(−0.649)
Std Dev 1.805***
(3.231)
Prior Earn Announcement −0.143***
(−14.405)
Analyst Forecast Dispersion 0.035⁎
(continued on next page)
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
371
The coefficient on Δ Log Non-MC Roadshow is insignificant in column (1) but positive and significant in column (2), suggesting that
private meetings with investors in non-money center high institutional ownership areas do not appear to systematically lead to
increases in ownership, but result in more buying and selling activity by these investors. Overall, the evidence in Table 6 is consistent
with managers using jet flights for private meetings that result in significantly greater trading activity by local institutional investors.
5.2. Institutional investor trading gains
While the Table 6 results suggest that local institutional investors trade in response to the roadshows, it is not clear whether such
trading is informed. Thus, we investigate whether local institutional investors benefit from jet flights to the MSA by examining
trading gains. Specifically, we test whether local institutional investors’ trading gains increase in roadshow flights to the MSA
unconditionally and in subsamples partitioned based on motivations to fly to meet with investors at different locations. We chose the
partitioning variables based on the results in Table 3, where proxies for complexity of information (Intangible) and news events (Prior
Earn Announcement and Debt/Equity Issue) are significant in the MC Roadshow regression and proxies for commitment to transparent
disclosure to select investors (Log Prior MC Roadshow and Log Prior Non-MC Roadshow) are significant in both MC Roadshow and Non-
MC Roadshow regressions. We expect trading gains to be concentrated in subsamples where local investors are likely to receive the
most benefit from a private meeting; i.e., prior to a security issuance, after earnings news, and when the information is complex. We
also expect that potential trading gains are lower when the meetings are part of regularly scheduled meetings between managers and
selected investors whom managers have committed to keeping informed about the firm; thus, trading gains should be larger for
infrequent or unexpected private meetings compared to frequent or expected meetings.
We estimate Trading Gains by multiplying the quarterly changes in local institutional investors’ holdings by the firm’s buy-andhold
size-adjusted returns over the subsequent quarter for each firm-calendar-quarter-MSA. Thus, Trading Gains is positive (negative)
when institutional investors trade in the correct (opposite) direction of future returns. Positive coefficients on the roadshow variables
(Δ Log MC Roadshow and Δ Log Non-MC Roadshow) would suggest that institutional investors who increase (decrease) their holdings
in the firm in response to private meetings are likely to do so in anticipation of higher (lower) future returns.24
Table 5 (continued)
Log analyst forecast
(1.870)
Δ Short Interest −0.153
(−0.542)
Debt/Equity Issue 0.015
(1.050)
Announcement Ret 0.003
(0.055)
Intercept −0.621***
(−7.465)
N 44,413
Adj R2 0.1290
*,**,*** Significantly different from zero at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
†, ††, ††† MC Roadshow (Non-MC Roadshow) significantly greater than MC Single (Non-MC Single) at the 0.10, 0.05, and 0.01 level, using twotailed
tests.
This table presents regressions that test the relation between analyst forecasting activity and both money center flights and non-moneycenter
flights to locations with high firm-specific ownership. The unit of observation is the three-day flight window (see Appendix A). The
dependent variable is Log Analyst Forecast, which is measured as the log of one plus the number of analyst EPS forecasts issued during the
three-day flight window. We identify three-day flight windows using indicator variables. MC Single is an indicator variable coded as 1 for
three-day window with flights to only one money center, and 0 otherwise. MC Roadshow is an indicator variable coded as 1 for three-day
window with flights to two or more money centers, and 0 otherwise. We define money centers as locations in the Boston, Chicago, New York,
and San Francisco MSAs. See Appendix A for more details. Non-MC Single is an indicator variable coded as 1 for three-day window with
flights to a single city where the firm has a high level of firm-specific institutional ownership, and 0 otherwise. Non-MC Roadshow is an
indicator variable coded as 1 for three-day window with flights to two or more cities where the firm has a high level of firm-specific
institutional ownership, and 0 otherwise. We define high firm-specific institutional ownership cities as any MSA with at least five institutional
owners or 1% ownership in the firm. Large City is an indicator variable coded as 1 if there is one flight to any of the next ten MSAs with
the most frequent number of flights in the three-day window, and 0 otherwise. Multiple Large Cities is an indicator variable coded as 1 if there
are flights to two or more cities in the next ten MSAs with the most frequent number of flights in the three-day window, and 0 otherwise. The
omitted category is three-day window with flights to other cities for which a firm has low or no institutional ownership. We include a control
for the number of analyst forecasts in the 90-day period prior to the flight window (Log Prior Analyst Forecast). Variable definitions are
provided in Appendix B. Standard errors clustered at both the firm- and quarter-level. All continuous variables are winsorized at the 1st and
99th percentiles.
24 Ideally, we would only measure trading gains for institutional investors that met privately with managers. However, this is not observable, and we must use all
institutional ownership in the MSA. If the private meetings are only with a small subset of local investors, it will add noise to this measure and weaken the power of the
tests.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
372
Table 6
Changes in institutional ownership at the MSA-level and flights to the MSA.
Δ Institutional ownership Abs (Δ Institutional ownership)
Δ Log MC Single 0.023 0.069***
(0.562) (3.487)
Δ Log MC Roadshow 0.053*** 0.054***
(2.814) (3.156)
Δ Log Non-MC Single 0.012 −0.001
(1.481) (−0.130)
Δ Log Non-MC Roadshow −0.002 0.016**
(−0.589) (2.344)
Δ Log Large City 0.003 −0.013
(0.217) (−0.982)
Δ Log Multiple Large Cities 0.024 −0.025
(1.049) (−1.443)
Debt/Equity Issue 0.002 0.001
(0.481) (0.229)
EP Ratio 0.106** −0.029
(2.361) (−0.572)
BM Ratio 0.002 0.000
(0.307) (0.048)
Leverage −0.015*** 0.011
(−2.619) (0.771)
Intangible 0.005 0.045***
(0.740) (3.322)
Δ Earn −0.141 −0.091
(−1.419) (−1.229)
Sales Growth −0.008 0.012
(−0.859) (1.332)
Log MVE −0.001 −0.067***
(−0.720) (−4.638)
Log Analyst Coverage 0.003 0.001
(0.940) (0.315)
Beta 0.003 −0.006
(0.845) (−0.780)
Std Dev −0.053 0.174
(−0.488) (0.466)
Returns 0.011 0.011⁎
(0.749) (1.706)
Share Turnover −0.005 0.215***
(−0.581) (3.337)
Δ Short Interest 1.661*** −0.147
(3.284) (−0.790)
Prior Earn Announcement −0.008 −0.009
(−1.292) (−1.142)
Announcement Ret −0.007 0.005
(−0.159) (0.154)
Analyst Forecast Dispersion −0.001 0.012
(−0.105) (1.606)
N 291,225 291,225
Adj R2 0.002 0.034
*,**,*** Significantly different from zero at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
†, ††, ††† MC Roadshow (Non-MC Roadshow) significantly greater than MC Single (Non-MC Single) at the 0.10, 0.05, and 0.01 level, using two-tailed
tests.
This table presents regression that tests the relation between quarterly changes in local institutional ownership and changes in the frequency of
flights and roadshows to the MSA. The regression is estimated at the Firm-Quarter-MSA level. In column (1), the dependent variable is Δ Institutional
Ownership, which is measured as the quarterly change in percentage of shares outstanding held by institutional investors located in the MSA. In
column (2), the dependent variable is Abs(ΔInstitutional Ownership), which is measured as the absolute change of Δ Institutional Ownership. Our
flight variables of interest are changes in quarterly flights. Δ Log MC Single is the quarterly change in log(1+number of 3-day flight windows with
flights to only one money center). Δ Log MC Roadshow is the quarterly change in log(1+number of 3-day flights windows with flights to two or
more money centers). Δ Log Non-MC Single is the quarterly change in log(1+number of 3-day flight windows with flights to a single city where the
firm has a high level of firm-specific institutional ownership). Δ Log Non-MC Roadshow is the quarterly change in log(1+number of 3-day flight
windows with flights to two or more cities where the firm has a high level of firm-specific institutional ownership). Δ Log Large City is the quarterly
change in log(1+number of 3-day flight windows that contain one flight to any of the next ten MSAs with the most frequent number of flights. Δ
Log Multiple Large Cities is the quarterly change in log(1+number of 3-day flight windows that contain flights to two or more of the next ten MSAs
with the most frequent number of flights).Variable definitions are provided in Appendix B. We include quarter fixed effect and cluster standard
errors at the MSA- and Firm-Quarter- level. All continuous variables are winsorized at the 1st and 99th percentiles.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
373
Table 7 presents the results of regressions of Trading Gains on the same set of variables as in Table 6, and an additional variable (BHARq) to
control for buy-and-hold size-adjusted returns over the current quarter (Bushee and Goodman, 2007). We use the same firm-quarter-MSA
sample as in Table 6 with the additional requirement of subsequent returns, which reduces the sample size to 251,625.
In the first column, we find that the coefficients on the roadshow variables are not significant. Thus, while local institutional
investor trading is greater, on average, when there are roadshows, there is no evidence that roadshows are unconditionally related to
greater trading gains.
In the other columns, we partition the sample based on variables that were significant determinants of roadshows in Table 3: (i)
whether the firm has a debt or equity issue in the subsequent month (Debt/Equity Issue), (ii) whether it is the month after the earnings
announcement (Prior Earn Announcement), (iii) whether the firm’s intangibles (Intangible) is in the top quartile in the quarter, and (iv)
whether there is only one or zero flights to the MSA in the prior quarter. We find a significant association between changes in MC
Roadshow and Trading Gains in the subsamples where complexity is high (Intangible in top quartile=1) and where private meetings
between managers and investors in the MSA are infrequent (prior flight<=1). There is also evidence of significantly larger trading
gains for MC Roadshow than MC Single before debt/equity issuances and in the month after the earnings announcement, but neither of
the coefficients are significantly different from zero on their own (though they are large in magnitude). Thus, there is no widespread
evidence of institutional investors being able to earn trading gains based on roadshow meetings, but there is some evidence that
trading gains exist when the firm’s information is more complex and prior private meetings between managers and investors were
infrequent.
Table 7
Regression of trading gains at the MSA-level on flights to the MSA.
Full sample Partitioning variable
Debt/equity issue Prior earn announcement Intangible in top quartile Prior flights to MSA<=1
1 0 1 0 1 0 1 0
Δ Log MC Single −0.026 −0.753 0.037 −1.722 0.106 −0.246 0.057 0.817 −0.299
(−0.08) (−0.290) (0.090) (−1.482) (0.271) (−0.340) (0.200) (0.74) (−1.11)
Δ Log MC Roadshow 0.404 2.704†† 0.201 2.476†† 0.233 0.753** 0.232 2.176*** 0.022%%%
(1.07) (0.990) (0.560) (1.298) (0.532) (2.190) (0.420) (3.70) (0.04)
Δ Log Non-MC Single −0.044 1.000 −0.132% 1.693 −0.154 −0.204 −0.009 0.068 −0.054
(−0.16) (1.400) (−0.460) (1.418) (−0.535) (−0.550) (−0.030) (0.18) (−0.16)
Δ Log Non-MC Roadshow 0.002 −0.172 0.025 −0.669 0.037 −0.168 0.068 0.199 −0.113
(0.02) (−0.440) (0.160) (−1.158) (0.256) (−0.640) (0.370) (0.99) (−0.53)
Δ Log Large City −0.138 −0.749 −0.111 0.357 −0.168 0.159 −0.168 −0.744 0.243
(−0.71) (−0.590) (−0.590) (0.879) (−0.729) (0.330) (−0.690) (−1.04) (0.87)
Δ Log Multiple Large Cities −0.267 5.532 −0.450 −1.881⁎ −0.186% −0.727 −0.213 −0.816 −0.074
(−0.68) (0.870) (−0.880) (−1.673) (−0.520) (−1.260) (−0.530) (−0.79) (−0.13)
BHAR 0.161 1.576 0.147 1.759 0.083 −0.815 0.287 0.044 1.210
(0.56) (1.400) (0.460) (1.317) (0.315) (−0.820) (1.090) (0.17) (0.66)
Debt/Equity Issue −0.001 0.076 −0.041 −0.086 −0.033 0.075 −0.525
(−0.02) (0.172) (−0.641) (−0.480) (−0.380) (1.01) (−1.49)
EP Ratio 0.676 −1.907 0.803 4.796 −0.030 8.386*** −0.484 −0.316 8.243
(0.59) (−0.550) (0.600) (1.130) (−0.033) (2.690) (−0.390) (−0.30) (0.95)
BM Ratio 0.051 0.891⁎ 0.013 −0.806⁎ 0.126 0.115 0.044 0.026 0.160
(0.32) (1.690) (0.080) (−1.805) (0.740) (0.300) (0.280) (0.16) (0.44)
Leverage −0.191 0.052 −0.194 0.546 −0.239 0.333 −0.245 −0.065 −1.343
(−0.55) (0.100) (−0.540) (0.362) (−0.682) (0.920) (−0.640) (−0.27) (−0.58)
Intangible 0.081 0.929 0.031 0.623 0.043 0.156 −0.662
(0.49) (1.170) (0.210) (0.735) (0.197) (1.13) (−0.67)
Δ Earn 0.417 10.208 −0.050 −5.072 0.540 −12.224** 2.384 −0.230 7.305
(0.19) (0.900) (−0.020) (−0.338) (0.250) (−2.000) (1.290) (−0.12) (0.90)
Sales Growth −0.059 −0.432 −0.035 0.891 −0.056 0.359 −0.202 −0.070 0.257
(−0.28) (−0.480) (−0.160) (0.716) (−0.284) (0.970) (−0.850) (−0.33) (0.45)
Log MVE 0.008 −0.256 0.024 0.341 0.008 0.009 0.013 0.001 0.055
(0.33) (−1.440) (0.900) (1.329) (0.304) (0.190) (0.400) (0.04) (0.50)
Log Analyst Coverage 0.050 0.249 0.035 −0.908 0.049 −0.022 0.072 0.062 −0.010
(0.89) (1.280) (0.570) (−1.274) (0.724) (−0.490) (1.000) (1.28) (−0.05)
Beta 0.145 0.092 0.136 −0.545 0.219⁎ 0.360 0.077 0.102 0.629
(1.38) (0.450) (1.170) (−1.169) (1.809) (1.000) (1.040) (1.07) (1.38)
Std Dev −10.081⁎ −7.392 −9.145 28.914 −13.244⁎ −6.988 −8.936⁎ −11.417⁎ −2.210
(−1.81) (−0.710) (−1.530) (1.315) (−1.932) (−0.420) (−1.670) (−1.73) (−0.14)
Returns −0.072 −0.834 −0.041 0.013 −0.085 0.090 −0.085 0.130 −1.871
(−0.61) (−0.810) (−0.390) (0.027) (−0.700) (0.590) (−0.680) (1.26) (−1.46)
Share Turnover 0.673 −1.911 0.798 2.754 0.614 −0.062 0.711 0.439 3.024
(1.50) (−1.630) (1.560) (0.959) (1.301) (−0.090) (1.450) (1.24) (1.40)
Δ Short Interest 8.991 27.769*** 8.064 7.999 8.552 11.172 9.116 9.123 2.414
(1.35) (3.180) (1.210) (0.351) (1.150) (0.950) (1.310) (1.46) (0.22)
(continued on next page)
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
374
6. Conclusion
We provide evidence that managers use corporate jets for private meetings with investors that are ex-ante unobservable and that
these meetings affect stock prices, trading volume, analyst forecasts, and local institutional investor trading. We proxy for unobservable
private meetings using roadshows, which are three-day windows during which the corporate jet flies to multiple money
centers or other high-ownership cities. We find that the frequency of roadshows is significantly associated with a number of proxies
for incentives to privately meet with investors, including the complexity of the firm’s information, investor demands for privileged
access, changes in the firm’s information environment, and a commitment to provide transparency to selected key investors. These
determinants do not explain flights to other locations with the same sign and significance, validating our roadshow variables as a
measure of unobservable private meeting.
We find significant positive abnormal absolute size-adjusted returns and abnormal share turnover during three-day money center
roadshow flight windows that are also significantly greater than windows with flights to only one money center. Furthermore,
roadshow flights to non-money-center cities that have a high local firm-specific institutional ownership are associated with greater
market reaction than flights to single high-ownership cities or cities with little or no institutional ownership in the firm. We also find
that signed size-adjusted returns are positive for both types of roadshows, suggesting that managers are more likely to undertake
roadshows when they believe the stock is undervalued (as opposed to managing bad news).
We also find that roadshows to money centers are associated with more analyst forecast activity, whereas roadshows to other
high-ownership cities are not. This finding further validates roadshows as a measure of private meetings by showing that flights to
places where most analysts are located are associated with more forecasting activity. Finally, roadshows are also associated with
greater absolute changes in local institutional ownership, but such institutional investor trading is only associated with trading gains
in certain contexts (more complex information and when private interactions between managers and investors in the MSA are
infrequent). Thus, private meetings facilitated by roadshows are associated with increased analyst forecast activity and local institutional
investor trading, but do not provide widespread informed trading opportunities to local investors.
A limitation of this study is that we measure the event date of private meetings with noise (i.e., not all roadshows are for private
meetings). Such noise likely contributes to the low economic magnitudes of the results. Additionally, only a small subset of the
market meets with managers; non-participants who are unaware of the meetings do not realize they are at an information disadvantage,
and may trade in the opposite direction of the participants. Thus, the magnitudes of our results represent a lower bound
on the effect that they likely have on the meeting participants. Overall, the statistical significance and consistency of the results
suggest that these meetings are an important information event for the participating investors, raising the regulatory question of
whether non-participants are significantly disadvantaged by the real-time unobservability of corporate jet flight patterns.
Table 7 (continued)
Full sample Partitioning variable
Debt/equity issue Prior earn announcement Intangible in top quartile Prior flights to MSA<=1
1 0 1 0 1 0 1 0
Prior Earn Announcement −0.148 −0.424 −0.145 0.161 −0.265 −0.246⁎ 0.495
(−1.05) (−0.930) (−1.120) (0.500) (−1.610) (−1.83) (0.71)
Announcement Ret 0.724 5.363 0.414 −0.715 0.865 1.110 0.664 −0.307 9.646**
(1.59) (1.250) (1.040) (−0.310) (1.468) (0.860) (1.050) (−0.50) (2.33)
Analyst Forecast Dispersion −0.094 1.162 −0.179 0.334 −0.091 −0.187 −0.054 −0.027 −0.881
(−0.39) (0.950) (−0.650) (0.300) (−0.380) (−0.430) (−0.220) (−0.10) (−0.53)
N 251,625 17,883 233,742 8361 243,264 61,654 189,971 225,916 25,709
Adj R2 0.001 0.003 0.001 0.004 0.001 0.001 0.001 0.001 0.002
*,**,*** Significantly different from zero at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
†, ††, ††† MC Roadshow (Non-MC Roadshow) significantly greater than MC Single (Non-MC Single) at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
%,%%,%%% Significantly different from coefficient to the left at the 0.10, 0.05, and 0.01 level, using two-tailed tests.
This table presents regressions that test the relation between local institutional investors’ trading gains and changes in frequency of flights and roadshows to the MSA.
The regressions are estimated at the Firm-Quarter-MSA level. The first column presents results for the full sample. The remainder of the table presents results for
several subsamples. The subsamples are partitioned based on (i) whether there is either a debt or equity issue in the subsequent month (Debt/Equity Issue=1), (ii)
whether the firm experienced an earnings announcement return in the prior month (Earn Announcement Month), (iii) whether the firm’s Intangible is in the top quartile
in the quarter, and (iv) whether flights to the MSA are infrequent (prior flights=<1). The dependent variable is Trading Gains, which is measured as the quarterly
change in percentage of shares outstanding held by institutional investors located in the MSA multiplied by the buy-and-hold size-adjusted returns over the subsequent
quarter. Our flight variables of interest are changes in quarterly flights. Δ Log MC Single is the quarterly change in log(1+number of 3-day flight windows with flights
to only one money center). Δ Log MC Roadshow is the quarterly change in log(1+number of 3-day flights windows with flights to two or more money centers). Δ Log
Non-MC Single is the quarterly change in log(1+number of 3-day flight windows with flights to a single city where the firm has a high level of firm-specific
institutional ownership). Δ Log Non-MC Roadshow is the quarterly change in log(1+number of 3-day flight windows with flights to two or more cities where the firm
has a high level of firm-specific institutional ownership). Δ Log Large City is the quarterly change in log(1+number of 3-day flight windows that contain one flight to
any of the next ten MSAs with the most frequent number of flights. Δ Log Multiple Large Cities is the quarterly change in log(1+number of 3-day flight windows that
contain flights to two or more of the next ten MSAs with the most frequent number of flights). We also include BHAR defined as buy-and-hold size-adjusted returns
over the current quarter as a control variable. Variable definitions are provided in Appendix B. We include quarter fixed effect and cluster standard errors at the MSAand
Firm-Quarter- level. All continuous variables are winsorized at the 1st and 99th percentiles.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
375
Appendix A. Example of flight window definitions
Summary of all flights for 3M Company for the last quarter of 2007:
Beg date End date NFLT Window
type
MC cities Non-MC cities Other cities
20071003 20071005 12 MC
Roadshow
Boston, Chicago
(2)
– Austin, Buffalo (2), Manchester NH, Park
Rapids (4), Quad Cities
20071008 20071010 12 MC Single Chicago Cincinnati Austin, Fayetteville AR, New Haven, Park
Rapids (7)
20071012 20071016 19 MC Single Chicago (2) Baltimore Austin, Charlotte, Park Rapids (13),
Toluca
20071017 20071019 6 Other – – Austin (2), Park Rapids (4)
20071022 20071024 8 Other – – Honolulu, New Haven, Park Rapids (4),
Sendai, Sydney
20071026 20071030 2 Other – – Park Rapids (2)
20071031 20071102 3 Non-MC
Single
– Los Angeles Lexington
20071105 20071107 6 MC Single Chicago – Anchorage, Austin, Decatur, Helsinki,
Shanghai
20071112 20071114 16 MC
Roadshow
Chicago (4),
New York (2)
Atlanta, Denver, Los
Angeles, San Jose
Anchorage, Charlotte, Columbus,
Phoenix, Salisbury, Salinas
20071119 20071121 3 Non-MC
Roadshow
– Miami, Milwaukee (2) –
20071123 20071127 9 MC
Roadshow
Boston, Chicago
(2)
– Austin, Bogotá, Park Rapids (3), Sao
Paolo
20071128 20071130 6 MC Single Boston – Austin, Helsinki, Park Rapids (2), Toluca
20071203 20071205 9 MC Single New York (2) Detroit Austin, London, Phoenix, Prescott,
Tucson, Wrocław
20071206 20071210 5 Other – – Austin (2), Fayetteville AR, London,
Manchester UK
20071211 20071213 7 MC Single New York (3) Baltimore Park Rapids (3)
20071217 20071219 10 MC Single Boston Atlanta, Philadelphia Austin (2), Greenville SC, Park Rapids (2),
Savannah, Springfield MO
20071220 20071224 7 MC
Roadshow
Boston, Chicago
(2)
Milwaukee Austin, Park Rapids (2)
20071226 20071228 1 MC Single Chicago – –
This appendix summarizes all of the flights for the 3M Company between October 3, 2007 and December 27, 2007. NFLT is the number of flights during each flight
window. “MC Cities” are money center MSAs (Boston, Chicago, New York, and San Francisco). “Non-MC Cities” are other high-ownership cities, defined as MSAs with
at least five institutional owners or 1% ownership in the firm. “Other Cities” have low or no institutional ownership in the firm. To form flight windows, we start by
identifying the first trading date with a flight to an MC City and then define that date and the next two trading dates as a money center window. Non-trading-day flights
are included with the flights on the next trading day (e.g., flights on Monday, October 8 also include flights on October 6–7). Any five-day windows are three-tradingday
windows that span a weekend. Once all money center windows are defined, we form three-day windows that do not include any money center flights. Window
Types are defined sequentially as follows: MC Roadshow=flights to more than one MC City; MC Single=flights to only one MC City, Non-MC Roadshow=flights to
more than one Non-MC City (and no flights to MC Cities); Non-MC Single=flights to only one Non-MC City (and no flights to MC Cities); Other=flights only to Other
Cities. 3M Company has a major manufacturing center in Austin, TX and a conference center in Park Rapids, MN, which accounts for the large number of flights to
those cities.
Appendix B. Variable definitions
Variables Definition
Indicator variables to identify flights during three day windows
MC Single = Indicator variable coded as 1 for three-day window with flights to only one money center, and 0
otherwise. We define money centers as locations in the Boston, Chicago, New York, and San
Francisco MSAs. See Appendix A for more details.
MC Roadshow = Indicator variable coded as 1 for three-day window with flights to two or more money centers, and 0
otherwise. We define money centers as locations in the Boston, Chicago, New York, and San
Francisco MSAs. See Appendix A for more details.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
376
Non-MC Single = Indicator variable coded as 1 for three-day window with flights to a single city where the firm has a
high level of firm-specific institutional ownership, and 0 otherwise. We define high firm-specific
institutional ownership cities as any MSA with at least five institutional owners or 1% ownership in
the firm.
Non-MC Roadshow = Indicator variable coded as 1 for three-day window with flights to two or more cities where the firm
has a high level of firm-specific institutional ownership, and 0 otherwise. We define high firmspecific
institutional ownership cities as any MSA with at least five institutional owners or 1%
ownership in the firm.
Large City = Indicator variable coded as 1 if there is one flight to any of the next ten MSAs with the most frequent
number of flights in the three-day window, and 0 otherwise.
Multiple Large Cities = Indicator variable coded as 1 if there are flights to two or more cities in the next ten MSAs with the
most frequent number of flights in the three-day window, and 0 otherwise.
Flights during monthly windows (including variables that are determinants of monthly flights)
No. MC Roadshow = Number of money center roadshows during the month
Log MC Roadshow = Log(1+the number of money center roadshows during the month)
No. Non-MC Roadshow = Number of non-money center roadshows during the month
Log Non-MC Roadshow = Log(1+the number of non-money center roadshows during the month)
No. Other Flight = Number of flights during the month to non-money-center cities that are not high firm-specific
institutional ownership cities (i.e. other flights)
Log Other Flight = Log(1+No. Other Flights)
Determinants of monthly flights
Intangible = Ratio of the firm’s intangible assets to total assets at the end of the prior fiscal quarter
Institutional Ownership = Percentage of shares outstanding held by institutional investors at the end of the prior calendar
quarter
Analyst Forecast
Dispersion
= Standard deviation of analyst EPS forecast scaled by the mean analyst EPS forecast in the prior
month
Δ Short Interest = Month-to-month change in the ratio of shares sold short to trading volume measured in the prior
month
Debt/Equity Issue = Indicator coded as 1 if there is either a debt or equity issuance in the subsequent month and 0
otherwise
Prior Earn
Announcement
= Indicator variable coded as 1 if the firm announces earnings in the prior month, and 0 otherwise
Announcement Ret = Value-weighted market-adjusted return for the three-day window (−1,+1) centered around the
earnings announcement date
Control variables
MVE = Firm’s market value of equity at the end of the prior fiscal quarter
Log MVE = Log(MVE)
BM Ratio = Ratio of the firm’s book value to market value of assets at the end of the prior fiscal quarter
Sales Growth = Quarter-on-quarter sales growth at the end of the prior fiscal quarter
Δ Earn = Quarter-on-quarter change in net income divided by total assets at the end of the prior fiscal quarter
EP Ratio = Basic earnings per share (excluding extraordinary items) scaled by price at the end of the prior fiscal
quarter
Leverage = Firm’s leverage ratio, calculated as the firm’s total debt to total assets at the end of the prior fiscal
quarter
Beta = Year-end beta of the firm’s equity from CRSP
Std Dev = Year-end standard deviation of the firm’s stock returns from CRSP
Analyst Coverage = Number of sell-side analysts that cover the firm based on analyst EPS forecasts in the prior month
Log Analyst Coverage = Log(1+Analyst Coverage)
Returns = Buy-and-hold market adjusted return for 12 months prior to 30 days before the start of the flight
month
Share Turnover = Average monthly share turnover for 12 months prior to 30 days before the start of the flight month
No. 8K = Number of 8-Ks issued during the flight month
Log 8K = Log(1+No. 8 K)
No. MC conference = Number of money center conference presentations
Log MC conference = Log(1+No. MC conference)
No. Non-MC
Conference
= Number of other conference presentations during the flight month.
Log Non-MC
Conference
= Log(1+No. Non-MC Conference)
Log Total Flight = Log(1+the number of total flights in the same month)
Log Prior Total Flight = Log(1+the number of total flights in the same month in the prior year)
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
377
Log Prior MC
Roadshow
= Log(1+the number of MC roadshows in the same month in the prior year)
Log Prior Non-MC
Roadshow
= Log(1+the number of Non-MC roadshows in the same month in the prior year)
Market reactions surrounding flight windows
Abnormal ASAR = Abnormal absolute value size-adjusted returns is measured as the difference between the three-day
absolute size-adjusted returns and the mean three-day absolute size-adjusted returns over a 90-day
estimation period divided by the standard deviation of the mean absolute size-adjusted returns over
the estimation period.
Abnormal Turnover = Abnormal share turnover is measured as the three-day volume divided by shares outstanding less the
average three-day turnover in the 90-day estimation period, multiplied by 100.
SAR = Three-day size-adjusted return during the flight window
Analyst forecasts analysis
Log Analyst Forecast = Log (1+the number of analyst EPS forecasts issued during the flight windows (0,+2))
Log Prior Analyst
Forecast
= Log (1+the number of analyst EPS forecasts prior to the flights (−90, −1))
Control variables tailored to market reactions and analyst forecasts tests
Log Prior MC Flight = Log(1+the number of money center flights in the prior 21 trading days)
Log Prior Non-MC
Flight
= Log(1+the number of non-money-center flights in the prior 21 trading days)
Changes in institutional ownership analysis
Δ Institutional
Ownership
= Quarterly change in percentage of shares outstanding held by institutional investors located in the
MSA.
Abs(Δ Institutional
Ownership)
= Absolute value of Δ Institutional Ownership
Δ Log MC Single = Quarterly change in log(1+No. MC Single)
Δ Log MC Roadshow = Quarterly change in log(1+No. MC Roadshow)
Δ Log Non-MC Single = Quarterly change in log(1+No. Non-MC Single)
Δ Log Non-MC
Roadshow
= Quarterly change in log(1+No. Non-MC Roadshow)
Δ Log Large City = Quarterly change in log(1+number of 3-day flight windows that contain one flight to any of the
next ten MSAs with the most frequent number of flights)
Δ Log Multiple Large
Cities
= Quarterly change in log(1+number of 3-day flight windows that contain flights to two or more of
the next ten MSAs with the most frequent number of flights)
Trading gains analysis
Trading gains = Quarterly change in percentage of shares outstanding held by institutional investors located in the
MSA multiplied by the buy-and-hold size-adjusted returns over the subsequent quarter.
BHAR = Buy-and-hold size-adjusted returns over the current quarter.
References
Andrews, A., Linn, S., Yi, H., 2009. Corporate governance and executive perquisites: evidence from the new SEC disclosure rules. Indiana University Working paper.
Beaver, W., 1968. The information content of annual earnings announcements. J. Accounting Res. 6, 67–92.
Bushee, B., Goodman, T., 2007. Which institutional investors trade based on private information about earnings and returns? J. Accounting Res. 45, 289–322.
Bushee, B., Jung, M., Miller, G., 2011. Conference presentations and disclosure milieu. J. Accounting Res. 49, 1163–1192.
Bushee, B., Jung, M., Miller, G., 2017. Do Investors benefit from selective access to management? J. Financial Rep. In-Press. Available at http://aaajournals.org/doi/
10.2308/jfir-51776.
Bushee, B., Matsumoto, D., Miller, G., 2003. Open versus closed conference calls: the determinants and effects of broadening access to disclosure. J. Accounting Econ.
34, 149–180.
Bushee, B., Miller, G., 2012. Investor relations, firm visibility, and investor following. Accounting Rev. 87, 867–897.
Cheng, Q., Du, F., Wang, X., Wang, Y., 2016. Seeing is believing: analysts’ corporate site visits. Rev. Accounting Stud. 21, 1245–1286.
Cooley, G., 2000., 2000. SEC Regulation FD: New Rules for Communicating with Analysts and Investors. Available at http://www.cooley.com/files/ALERT_SEC_Reg_
FD.pdf .
Cready, W., Hurtt, D., 2002. Assessing investor response to information events using return and volume metrics. Accounting Rev 77, 891–909.
Green, T.C., Jame, R., Markov, S., Subasi, M., 2014a. Broker-hosted investor conferences. J. Accounting Econ. 58, 142–166.
Green, T.C., Jame, R., Markov, S., Subasi, M., 2014b. Access to management and the informativeness of analyst research. J. Financial Econ. 114, 239–255.
Grinstein, Y., Weinbaum, D., Yehuda, N., 2017. The economic consequences of perk disclosure. Contemp. Accounting Res. 34 (4), 1812–1842.
Kirk, M., Markov, S., 2016. Come on over: analyst/investor days as a disclosure medium. Accounting Rev. 91, 1725–1750.
Koch, A., Lefanowicz, C., Robinson, J., 2013. Regulation FD: a review and synthesis of the academic literature. Accounting Horiz. 27, 619–646.
Landsman, W., Maydew, E., 2002. Has the information content of quarterly earnings announcements declined in the past three decades? J. Accounting Res. 40,
797–808.
Lang, M., Lundholm, R., 2000. Voluntary disclosure and equity offerings: reducing information asymmetry or hyping the stock? Contemp. Accounting Res. 17,
623–662.
Lee, L.F., Lowry, M., Shu, S., 2018. Are all perks solely perks? Evidence from corporate jets. J. Corporate Finance 48, 460–473.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
378
Maremont, M., McGinty, T., 2011. For the highest fliers, new scrutiny. Wall Street J.
Rajan, R., Wulf, J., 2006. Are perks purely managerial excess? J. Financial Econ. 79, 1–33.
Ryan, T., Jacobs, C., 2008. Using Investor Relations to Maximize Equity Valuation, Kindle Edition. (Wiley Finance).
SEC v. Siebel Systems, et al., 2005. United States District Court for the Southern District of New York, Memorandum Decision and Order, August 31.
Shroff, N., Sun, A., White, H., Zhang, W., 2013. Voluntary disclosure and information asymmetry: Evidence from the 2005 securities offering reform. J. of Account.
Res. 51, 1299–1345.
Sidhu, B., Smith, T., Whaley, R., Willis, R., 2008. Regulation fair disclosure and the cost of adverse selection. J. Accounting Res. 46, 697–728.
Solomon, D., Soltes, E., 2015. What are we meeting for? The consequences of private meetings with investors. J. Law Econ. 58, 325–355.
Soltes, E., 2014. Private interaction between firm management and sell-side analysts. J. Accounting Res. 52, 245–272.
Subasi, M., 2014. Investor Conferences and Institutional Trading in Takeover Targets. Working paper. University of Texas at Dallas.
Subramaniam, C., 1995. Detecting information content of corporate announcements using variance increases: a methodological study. J. Accounting Auditing Finance
(Fall) 415–30.
Tasker, S., 1998. Bridging the information gap: quarterly conference call as a medium for voluntary disclosure. Rev. Accounting Stud. 3, 137–167.
Yermack, D., 2006. Flights of fancy: corporate jets, CEO perquisites, and inferior shareholder returns. J. Financial Econ. 80 (1), 211–242.
Yermack, D., 2014. Tailspotting: identifying and profiting from CEO vacation trips. J. Financial Econ. 113, 252–269.
B.J. Bushee et al. Journal of Accounting and Economics 65 (2018) 358–379
379