Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Leases and Over-investment†
Tim Eaton Miami University
Craig Nichols Syracuse University [email protected]
James Wahlen
James R. Hodge Chair of Excellence Indiana University
Matthew Wieland Miami University
April 2018
†We thank Brian Bratten, Brian Callahan, Amanda Convery, Joe Comprix, Dave Harris, Shyam Sunder,
Jake Thomas, an anonymous reviewer for the 2017 AAA annual meeting, and workshop participants at Binghamton University, Miami University, Syracuse University, the University of Kentucky, and conference participants at Yale University for helpful comments and suggestions. All remaining errors are our own. Professor Wahlen gratefully acknowledges support from the James R. Hodge Chair of Excellence.
Leases and Over-investment
Abstract What economic consequences arise (if any) if firms overinvest in leased assets? We estimate proxies for over-investments in leased assets using positive residuals from an expected lease change model. We find that, compared to capital expenditures, over-investments in leased assets are less sensitive to free cash flows, and are not asymmetrically sensitive to positive free cash flows. Thus, leasing appears to be a mechanism for over-investment even in the absence of free cash flows. In examining economic consequences, we predict and find that overinvestments in leased assets trigger increasing future sales growth but declining future earnings growth, which persists for as long as three years. We also find a negative relation with contemporaneous stock returns, suggesting investors view over-investments in leases as value destructive, on average, even after controlling for their effects on sample firms’ discount rates. Finally, despite these negative earnings and returns consequences, we find that over-investments in leased assets are associated with higher future CEO compensation, arising primarily from future sales growth. Our results should inform boards, investors, and researchers interested in corporate governance. Key Words: Leases, Over-investment, Corporate Governance, Compensation, Sales and Earnings Growth, Stock Returns Data Availability: Data are available from sources identified in the paper
1
Leases and Over-investment
1. Introduction
It is a well-accepted empirical finding that some managers over-invest in acquisitions and capital
expenditures to drive firm growth (e.g., Lewellen et al. 1985; Lang et al. 1991; Richardson 2006; Biddle
and Hilary 2006).1 Consistent with Jensen (1986), managers often choose to invest free cash flows
(rather than distribute them to investors) by acquiring other companies or expanding physical plant,
even if such investments are negative net present value projects. The large body of prior literature on
over-investments has not examined potential over-investments in leased assets, even though they
surpass $1.5 trillion among publicly traded U.S. companies (U.S. Chamber of Commerce 2012). We
predict that over-investment also occurs through leases because initiating a lease does not require
issuing debt or equity securities, requires little upfront cash outflow, is generally accounted for as an
operating activity (i.e., rent expense), and typically does not trigger balance sheet recognition of assets
or liabilities. If firms over-invest in leased assets, what economic consequences arise (if any)?
Our results reveal that over-investments in operating leased assets trigger increasing future
sales growth but declining future earnings growth, persisting for up to three years. We also find that
overinvestments are negatively associated with contemporaneous stock returns, indicating the capital
markets view them as value destructive. Finally, we predict and find that overinvestments in leased
assets are positively associated with future growth in CEO compensation that is tied to future sales
growth, suggesting managers make these value-destroying investments in response to sales-growth-
based compensation incentives.
1 It is well known that managers of publicly traded companies have incentives that do not always align perfectly with shareholders’ incentives (Berle and Means 1932; Jensen and Meckling 1976; and many others). One manifestation of this agency conflict is a strong preference by some managers to run larger firms, which may give rise to over-investment (Baumol 1959; Marris 1964; Williamson 1964; Donaldson 1984; Jensen 1986, 1993; Stein 2003). Over-investment has been the subject of formal models by Stulz (1990), Harris and Raviv (1990), Hart and Moore (1995), and Zwiebel (1996). These models generally assume that managers enjoy private benefits of control, which are proportional to investment (Hart and Moore 1995) or the gross output (i.e., revenues) of the firm (e.g., Stulz 1990).
2
To conduct our empirical analyses, we examine 11,434 firm-year observations from 2000 to
2015 with operating lease footnote data in Compustat and for which the undiscounted future lease
commitments exceed five percent of total assets.2 We adapt Richardson’s (2006) model of investments
to estimate expected investments in leased assets. We first regress changes in operating lease assets on
firm characteristics designed to capture economic determinants of leases, using proxies for growth
opportunities, cash constraints, leverage, and lagged lease changes. The residual lease changes from
this regression should therefore reflect suboptimal amounts of investments in leases. We use positive
residuals from our model as proxies for over-investments in operating lease assets.
We then test the economic consequences of over-investments. Theoretical models of over-
investment (e.g., Stulz 1990) predict that managers are primarily focused on the gross output of the firm
(i.e., sales), whereas shareholders are more focused on wealth creation (i.e., earnings). Because new
investment in asset growth (rather than replacement of expiring assets) should increase a firm’s
productive capacity, we predict that over-investment in leased assets will drive sales increases in the
next period. However, over-investment is value-destructive (by definition), and should lead to lower
future earnings growth. Therefore, in our tests, we predict over-investments in leased assets will have a
positive association with future sales growth, but a negative association with future earnings growth.
Intuitively, such a pattern indicates that the costs of the investment outweigh the benefits.
We use both the expected and the unexpected amounts of investments in leases in predictive
models of one-year ahead growth in sales and earnings, while controlling for other determinants,
including current period growth in sales and earnings. We find that unexpected investments in leases
relate positively to future sales growth, but negatively to future earnings growth. These results fit the
profile of over-investment. In contrast, expected lease changes relate positively to future sales growth,
2 We focus on operating leases because they are more common and a more important source of financing than capitalized leases. In our sample, discounted minimum future operating lease payments amount to 18.6 percent of total assets. Only 36.4 percent of our sample firms also have capital leases, and these obligations comprise only 1.7 percent of total assets.
3
but have a neutral relation with future earnings growth. That is, expected lease changes do not reflect
over-investment. We also divide the sample based on the sign of the unexpected lease change. The
results verify an over-investment effect by showing that positive unexpected lease changes are
associated with negative future earnings changes.
These results are limited because they only examine sales and earnings changes one period
ahead; most leases have multi-period effects. Periodic lease expense begins to accrue at inception, yet
the leased asset may not generate sales for weeks or months, and may not reach its full revenue- and
earnings-generating potential until well after the first year. To address this issue, we perform two sets of
tests. First, we expand the window to three years to capture more of the costs and benefits realized
over the lease life. The over-investment effects persist for years +2 and +3. Unexpected changes in
leases have a positive association with future sales growth for each of the next three years, and a
negative relation with future earnings growth over each of the next three years. Again, the over-
investment effects are concentrated in unexpected increases in leases; expected lease changes have a
significant positive association with sales growth over the next three years, yet no significant relation
with future earnings growth. Moreover, the over-investment effects are particularly concentrated within
the positive unexpected increases in leases: the positive association with future sales growth diminishes
in years +2 and +3, whereas the negative association with earnings growth remains significant over each
of the next three years.
Second, we examine the relation between over-investments in leased assets and
contemporaneous stock returns, which reflect revisions in investors’ expectations of the present value
of future cash flows. If stock prices reflect lease information, contemporaneous stock returns should be
decreasing in over-investments in leases. Our results are consistent with this prediction. However,
because leases increase a firm’s leverage, they will increase the firm’s cost of equity capital (Bratten et
al. 2013; Dhaliwal et al. 2011). Lease increases should, therefore, be related to contemporaneous
4
increases in discount rates, which drive stock prices down, holding expected cash flows constant (e.g.,
Campbell and Shiller 1988; Vuolteenaho 2001; Campbell and Vuolteenaho 2004). Because unexpected
lease increases relate negatively to future earnings growth over the next three years, they are also likely
negatively associated with future cash flows. Thus, a negative relation between over-investments in
leases and contemporaneous stock returns could be driven by lower expected future cash flows (an
over-investment effect), higher discount rates (a leverage effect), or both. To disentangle these effects,
we adapt the Penman and Yehuda (2016) approach to control for changes in expected returns during
the period. After controlling for the discount rate effect, the relation between unexpected lease changes
and contemporaneous stock returns becomes more strongly negative. Overall, our evidence suggests
that when managers over-invest in leases, it leads to faster future sales growth over three years, slower
future earnings growth over three years, and lower contemporaneous stock returns.
For comparative analyses, we also estimate Richardson’s (2006) model for expected
investments in capital expenditures. We find the consequences of over-investing in leases are
incremental to the consequences of over-investing in capital expenditures (CapEx). Specifically, future
sales growth is increasing in over-investments in leases as well as in CapEx, while future earnings growth
is decreasing in both forms of over-investment. We find similar long-run consequences from over-
investments in capital expenditures, which are positively related to three-year sales growth, but strongly
negatively related to three-year earnings growth. The relation between unexpected investments in
capital expenditures and contemporaneous stock returns is also strongly negative.
We next examine how unexpected lease changes relate to unexpected CapEx. We find that firms
in the highest (lowest) quintile of unexpected lease changes also have high (low) unexpected CapEx.
Thus, leasing and capital expenditures are not simply substitute mechanisms for over-investment. To
explore how leases and CapEx relate to free cash flows, we sort firms into quintiles based on current
period free cash flows. We find that over-investment in CapEx is more sensitive to free cash flows;
5
unexpected lease changes have a weaker association with free cash flows. In regression tests, we find
that cash flow sensitivity of unexpected lease changes does not differ with the sign of free cash flows. In
contrast, unexpected CapEx is highly sensitive to free cash flows when free cash flows are positive. This
result is consistent with Jensen’s (1993) free cash flow hypothesis for CapEx, but also shows that
managers can over-invest through leases even in the absence of positive free cash flows.
Why would managers over-invest? In general, over-investment can enlarge the firm’s economic
footprint and the manager’s power. Growth in assets and revenues generated by over-investment can
heighten the visibility and reputation of the firm and the manager; create more perquisites to dispense
or consume; and give rise to other intangible benefits that managers can derive from a larger firm.
Because these benefits are difficult to isolate and measure, we instead explore the impact of over-
investment on managers’ compensation. Studies of executive pay indicate that firms often create
explicit incentives to grow by linking compensation to revenues. Murphy (1998) notes that
compensation committees often set base salaries for CEOs through a benchmarking process, selecting
peer firms based on size, measured by the level of sales. In addition, Murphy (1998) concludes, “each
dollar increase in base salary has positive repercussions on many other compensation components,” and
the primary determinant of base salary is company size. Core, Holthausen, and Larcker (1999) find that
sales is the statistically strongest cross-sectional determinant of executive compensation, exceeding
investment opportunities, return on assets, stock returns, and various board, governance, and
ownership structure factors. More recently, Huang, Marquardt, and Zhang (2015) report that sales has
become the most frequently used explicit performance measure in executive incentive plans.
Our final analyses examine whether managers experience positive compensation consequences
from over-investments in leases. We use the Core, Holthausen and Larcker (1999) model (absent their
corporate governance and ownership structure variables) and obtain similar results. We find that, for
the firms in our sample, CEO compensation is increasing in profitability, investment opportunities, and
6
stock returns, and most importantly, strongly increasing in lagged sales. We also find that CEOs of firms
that over-invest in leases experience compensation increasing in sales growth associated with over-
investments in leases. Even within the subsample of firms we classify as over-investing in leases, CEOs
receive strongly increasing compensation as a function of sales. By contrast, we find that CEOs of firms
that appear to over-invest in capital expenditures receive significantly higher total compensation, but
experience compensation decreasing in sales growth associated with over-investments in capital
expenditures.
Our study should interest managers, boards, and researchers interested in corporate
governance because our results should increase awareness of the potential for over-investment among
companies engaged in leasing. We provide evidence on the implications of over-investments in leased
assets for future firm performance and CEO compensation. Managers may over-invest in leases to
respond to the sales growth incentives set by the board and the compensation committee, even though
over-investments can destroy value. Our results suggest that compensation committees should consider
placing less emphasis on sales, not more, as has been the recent trend (Huang et al. 2015).
For researchers, we extend the lease literature by providing novel evidence on economic
consequences associated with over-investments in off-balance sheet leases. Additionally, we contribute
to the research on agency problems by showing that the potential over-investment problems extend
beyond free cash flows. Finally, we develop tests of the consequences that current period over-
investments trigger for future sales, earnings, returns, and compensation that future researchers can
use to test the factors influencing over-investment.
In the remainder of this paper, we discuss related literature and our predictions in section 2. We
outline our sample selection, variable measurement, and research design in section 3. We discuss
results of our primary tests and supplemental analyses in section 4. We conclude in section 5.
7
2. Related literature and predictions
First, we discuss the literature on agency problems, compensation, and over-investment, and
then we discuss the accounting literature on the implications of leases for risk, credit assessments, and
the cost of capital.3 Second, we develop our predictions for the economic consequences associated with
over-investments in right-of-use assets and obligations.
2.1 Agency problems and compensation
Starting with Berle and Means (1932) and advanced by Jensen and Meckling (1976) and others,
it has long been recognized that the incentives of shareholders and managers of public firms do not
perfectly align. Agency conflicts can manifest in managers having an “excessive taste for running large
firms, as opposed to simply profitable ones” (Stein 2003) and over-investing (Baumol 1959; Marris 1964;
Williamson 1964; Donaldson 1984; and Jensen 1986, 1993). Empire-building and over-investment have
been the subject of formal models by Stulz (1990), Harris, and Raviv (1990), Hart and Moore (1995), and
Zwiebel (1996). These models generally assume that managers enjoy private benefits of control, which
are proportional to investment (Hart and Moore 1995) or the gross output of the firm (Stulz 1990).
Managers’ compensation contracts create explicit rewards for sales and sales growth, which can
exacerbate the incentives for over-investment. The literature on executive pay indicates that firms
create incentives to grow by linking manager compensation to sales. Surveying this literature, Murphy
(1998) notes that a typical compensation committee establishes the base salary for the CEO through a
benchmarking process, selecting peer firms based on sales levels. Base salaries represent the fixed
component of the compensation package. Thus, risk-averse executives will prefer a dollar increase in
salary over a dollar increase in bonus or incentive-based pay. Moreover, compensation committees link
other components of the compensation package to the base salary level. As Murphy (1998) points out,
compensation packages typically express the potential bonus pool as a percentage of base salary, and
3 This section relies on reviews in Stein (2003), Murphy (1998), Lipe (2001), and Spencer and Webb (2015).
8
often set option grants as a multiple of base salary. Murphy (1998) concludes, “each dollar increase in
base salary has positive repercussions on many other compensation components,” and the primary
determinant of base salary is company size as measured by the level of sales.
Empirical studies on executive pay have established the strong influence of sales in the
compensation practices of firms. In summarizing this literature, Murphy (1998) notes that a standard
result across studies is a cross-sectional elasticity of compensation to sales of about 0.3. Thus, a firm
with 10 percent higher sales will pay its executives an average of 3 percent more. Murphy (1985) also
notes that this influence of sales is so strong that measures of net performance (return on equity,
profits) play “at best a minor role.” The influence of sales on executive pay even outweighs stock
returns. Murphy (1985) finds that, controlling for size, a firm with a 10 percent stock return will pay its
executives 1 percent more than a firm with zero percent return. In contrast, holding stock return
performance constant, a firm with 10 percent higher sales will pay its executives 2.8 percent more. As
Baker, Jensen, and Murphy (1988) conclude, these results suggest that CEOs can increase their pay by
increasing firm sales, even when the increase in sales reduces the firm’s market value.
Similarly, Core, Holthausen, and Larcker (1999) find that sales is the statistically strongest cross-
sectional determinant of executive compensation, exceeding investment opportunities, return on
assets, stock returns, and various board, governance, and ownership structure factors. More recently,
Huang, Marquardt, and Zhang (2015) report that sales has become the most frequently used explicit
performance measure in executive compensation plans. They find that the proportion of firms using
sales as a performance measure increased from 25 percent in 2001 to 34 percent in 2010, and that the
sensitivity of executive pay to firm sales has increased over this period. Moreover, Huang et al (2015)
find that the sensitivity of pay to sales remains significant after controlling for earnings and returns.
2.2 Over-investment
9
Studies of over-investment typically focus on acquisitions and the sensitivity of capital
expenditures to free cash flows. Lewellen, Loderer, and Rosenfeld (1985) find that acquiring firms tend
to overpay more when the acquiring firm’s top management has a small equity ownership, and thus the
interests of managers and shareholders diverge. Lang, Stulz, and Walkling (1991) find the acquirers with
high free cash flows and low investment opportunities are more likely to overpay, consistent with
Jensen’s (1986) free cash flow hypothesis.4
Although it is clear that firms with greater cash on hand and free cash flows invest more, this
could be because external financing is costly (Myers and Majluf 1984), because investment
opportunities are not independent of past performance, and/or because managers prefer to manage
large firms (e.g., Jensen 1986, 1993). Although over-investment is a common explanation (Jensen 1986,
1993), the evidence linking investment/free-cash-flow sensitivity to agency conflict is sparse (Stein
2003), in part due to difficulties in measuring over-investment.5
Several papers examine over-investment in the accounting literature. Richardson (2006)
develops a measure of over-investment by regressing new capital expenditures on proxies for growth
opportunities, cash constraints, and leverage. Using his measure, he finds that firms with positive free
cash flows over-invest on average. Biddle and Hilary (2006) measure suboptimal investment as the
residual from regressing capital expenditures on lagged sales. They find lower degrees of suboptimal
investment for firms with higher financial reporting quality. Biddle, Hilary, and Verdi (2009) extend these
results to show that better financial reporting quality is associated with lower degrees of both over- and
4 Additional evidence of overpaying for acquisitions and wasteful spending related to agency conflicts can be found in Harford (1999), Opler, et al. (1999, 2001), and Blanchard, Lopez-de-Silanes, and Shleifer (1994). Various other studies also document a positive sensitivity of investment to free cash flows, including Schaller (1993), Bond and Meghir (1994), Calomiris and Hubbard (1995), Chirinko (1995), Gilchrist and Himmelberg (1995), and Lang, Ofek, and Stulz (1996), among many others. 5 Another stream of literature examines the role of leases in capital structure and financing decisions. For example, leases are an attractive source of capital for firms in financial distress (Realdon 2006) or with high costs of external funds (Sharpe and Nguyen 1995). This research examines whether firms use leases to complement existing debt (Lewis and Schallheim 1992; Schallheim, Wells, and Whitby 2013) or substitute for other forms of financing (Marston and Harris 1988; Krishnan and Moyer 1994; Yan 2006).
10
under-investment. However, it is unclear whether the results extend to investments in leases. For
example, Richardson (2006) focuses on firms with positive free cash flows. Leased assets, in contrast,
require little cash outflow at lease inception. For firms interested in investing excess free cash flows,
leases are probably not the top choice.
2.3 Economic consequences associated with leases
Studies in accounting examine the economic consequences of leases as a form of leverage, but
they focus on risk and costs of capital rather than over-investment (see Lipe 2001, and Spencer and
Webb 2015 for reviews). Various studies find that leases increase stock return volatility (Imhoff et al
1993; Ely 1995). Although studies examining market beta provide mixed results (e.g., Bowman 1980),
Bratten et al (2013) and Dhaliwal et al (2011) show that leases increase the cost of equity.
Leases also help predict distress (Elam 1975), and increase credit risk and the cost of debt. Kraft
(2015) finds that Moody’s analysts incorporate leases into their quantitative measures of financial
position, as well as in their qualitative assessments of credit-worthiness that shape credit ratings.
Similarly, Sengupta and Wang (2011) find that ratings reflect lease information. Credit ratings impact the
cost of debt, but Altamuro et al (2014) find that operating leases impact loan spreads even in the
absence of credit ratings. Chu et al. (2008) find that loan spreads increase in operating leases, although
they suggest the adjustment does not incorporate the full amount of the lease obligation. Bratten et al.
(2013) and Dhaliwal et al. (2011) also find that lease increase the costs of debt capital.
Although prior research documents a number of significant economic consequences associated
with leasing, the existing evidence does not examine whether leasing is associated with over-
investment.6 The agency costs associated with leased assets can be severe. Leases commit the firm to
6 In an unpublished working paper, Ge (2006) documents a negative relation between changes in off-balance sheet leases and levels of return on assets next period. Ge’s (2006) lease measure is somewhat limited as it does not include expected lease payments beyond five years because this item was unavailable in Compustat for the majority of her sample period. Moreover, Ge (2006) is careful to point out that the negative relation between current period lease changes and future return on assets need not reflect over-investment, but may simply reflect declining marginal rates of productivity. A negative relation between
11
pay out future cash flows for lease service rather than distributing them to shareholders. In contrast,
Jensen (1993) argues that issuing debt can be an effective safeguard against over-investment if the firm
pays out the debt proceeds to equity holders because this commits the firm to use future free cash
flows for debt service rather than over-investment. In this sense, leasing has a more negative effect for
equity holders: leases do not give rise to upfront cash proceeds that can be paid out to equity holders,
yet leases still commit the firm to use future free cash flows for lease service rather than payments to
equity holders.
2.4 Predictions: Over-investments in leases lead to higher sales changes and greater CEO
compensation but lower earnings changes and stock returns
Managers that derive benefit from the gross output of the firm (i.e., sales) rather than only the
net output (i.e., earnings) may have incentives to over-invest (Stulz 1990). Managers may over-invest in
leased assets to expand the productive capacity and output of the firm. An increase in leased assets in
the current period should therefore drive an increase in future sales.7
Over-investment occurs when a company takes projects for which the economic costs, including
the cost of capital, outweigh the economic benefits. Costs of capital play a critical role in any analysis of
net present value. Costs of capital are inherently difficult to measure, providing archival researchers a
significant obstacle in cleanly identifying whether an investment has a positive or negative net present
value. This is less of a concern in our setting because contractual lease payments give rise to lease
expense, which represents the agreed-upon payment between the firm and lessor for use of the capital.
As a result, the reported earnings number should reflect the benefits (stemming from sales increases)
net of the incremental costs (lease expense and other operating expenses) from investments in leased
current investment and future return on assets is expected, even if a company only invests in positive net present value projects. Thus, the results in Ge (2006) leave open the issue of over-investment in leased assets. 7 Increases in leased assets that reflect the renewal of expiring leases would not generally create new streams of sales, nor would they necessarily represent over-investment. Although we cannot identify the specific component of lease changes that capture lease renewals, we emphasize that these lease renewals will bias against our results.
12
assets. If firms over-invest in leases, the incremental costs should outweigh the benefits, causing future
earnings to be lower. We therefore focus on future earnings changes to test the effect of the investment
in leased assets. In the same vein, when the market observes the firm changing its investments in
leased assets and understands the implications for future earnings, then we would expect to observe a
negative relation with returns.
If managers over-invest in leased assets primarily to drive sales growth, and if those managers
receive compensation that is a function of sales growth, then we predict over-investments in leased
assets will also be associated with greater compensation for managers, despite the negative effect of
over-investments on earnings and stock returns.
To summarize, we predict that firms over-invest in off-balance sheet leased assets. Over-
investments in leases trigger higher future sales growth but lower future earnings growth. Further, we
predict lower expected future earnings growth associated with over-investments in leases will trigger
lower stock returns. We also predict that higher future sales growth arising from over-investments in
leases will be associated with higher pay for the managers, despite lower future earnings and returns.
We also examine the extent to which the economic consequences associated with over-investments in
leases exacerbate the over-investment problem with free cash flows and capital expenditures.
3. Research Design and Sample
In this section, we describe key features of our research design, beginning with how we measure
the lease obligation. We also describe our sample selection criteria. We then discuss how we estimate
the expected change in the lease obligation and the residual.
3.1 Measuring the Lease Obligation
We develop a measure of the lease obligation (which is approximately equivalent to the value of
leased assets) using the minimum lease commitments reported in Compustat (MRC1, MRC2, MRC3,
MRC4, MRC5, and MRCTA). The lease obligation variable (Lease) consists of two parts: (1) the minimum
13
future payments and (2) the discount rate. We begin with the minimum future lease payments for the
first 5 years. We then spread any remaining future commitments thereafter (MRCTA) over years +6 and
beyond on a straight-line basis by approximating the number of future periods represented in that
variable by dividing MRCTA by MRC5, the year +5 lease obligation. We cap the stream of future lease
payments at a maximum of 15 years, placing any leftover amount into year +15. For the discount rate,
we measure a yearly interest rate for each firm by dividing interest expense (XINT) by the average long-
term debt (DLTT). We then identify the firm-specific discount rate as the median interest rate for that
firm over the sample period. Using a firm-specific yet time-invariant discount rate provides assurance
that changes in future commitments cause changes in the lease obligation rather than changes in the
discount rate. We discount each of the minimum future lease payments using the firm-specific discount
rate.8 In Appendix A we illustrate the lease obligation computation using the information from Skywest,
Inc.’s 2012 Form 10-K. Skywest disclosed their estimate of the present value of the lease obligation, $1.8
billion, which approximates our estimate of the lease obligation variable ($1.8 billion).
3.2 Selecting the Sample
We begin with all firms in the intersection of Compustat and CRSP from 2000 to 2015.9 Because
of their unique accounting and economic characteristics, we eliminate firms in financial services
industries (SIC codes between 6000 and 6999). We also require firms to have price greater than $1 and a
market value of equity greater than $50 million. To identify firms for which leasing is important, we only
include firms for which the sum of undiscounted lease obligations exceeds five percent of total assets.10
Our final sample consists of 11,434 firm-year observations. We mitigate potential effects of outliers by
winsorizing the variables in the models at the top and bottom percentile.
3.3 Excluding capital leases
8 Our results are robust to this design choice. See section 4.4 for more information. 9 Our sample begins in 2000, when Compustat began providing the ‘after 5 year’ data for future contingent lease liabilities. 10 Our results are robust to this design choice. See section 4.4 for more information.
14
In identifying firms for which leases are important, we focus on operating leases rather than
capitalized leases. First, capital leases are far less common; only 36.4% of our sample firms also have
capitalized lease obligations. Applying our sample selection criteria would yield only 663 firm-year
observations with capital leases greater than 5 percent of total assets from 2000 to 2015 (only 6 percent
of the observations with operating leases in our sample). Moreover, among firms with capital leases,
operating leases are a much more important source of capital: capital lease obligations amount to only
1.7 percent of total assets, whereas discounted minimum future operating lease payments amount to
18.6 percent of total assets. Second, under current accounting rules, capital leases have a stronger
income-reducing impact on earnings in early years of the lease contract relative to leases. This occurs
because in the early years of a capital lease, the sum of (1) interest on the lease obligation and (2)
depreciation on the leased asset will amount to a larger total expense than the corresponding rent
expense for operating leases (see Lipe (2001) for additional discussion). Because operating leases have a
weaker income-reducing impact on earnings in early years of the lease, they provide a more
conservative test of our prediction of the impact of over-investment on earnings.
3.4 Descriptive statistics
Table 1 provides descriptive statistics of the discounted lease obligations in dollar amounts and
as percentages of total assets by year and industry. The mean lease obligation grows over the sample
period from $256.4 million in 2001 to $631.6 million in 2015. In most sample years, leases are roughly
between 18 and 20 percent of total assets. Not surprisingly, Panel B reveals that our sample contains the
greatest numbers of firm-year observations from the Services, Retail, Durable Manufacturers, and
Transportation industries. Firms in the Transportation and Retail industries rely the most heavily on
financing from operating leases, averaging $1,076.9 million and $903.7 million, respectively.
3.5 Expected Investments in Leased Assets
15
Some portion of new leases represent renewals of old leases that are expiring. Other new leases
arise in the pursuit of profitable growth opportunities. To estimate expected investments in leased
assets for maintenance and growth, we adapt Richardson’s (2006) investment model to leases. We
estimate the following model to determine expected investments in leases, and then estimate over-
investment (or under-investment) with the residuals:
𝐿𝑒𝑎𝑠𝑒_𝑐ℎ𝑎𝑛𝑔𝑒𝑖,𝑡 = 𝛼 + 𝛽1𝑉/𝑃𝑖,𝑡−1 + 𝛽2𝐿𝐸𝑉𝑖,𝑡−1 + 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1 + 𝛽4𝐴𝐺𝐸𝑖,𝑡−1 + 𝛽5𝑆𝐼𝑍𝐸𝑖,𝑡−1 +
𝛽4𝑅𝐸𝑇𝑈𝑅𝑁𝑖,𝑡−1 + 𝛽4𝐿𝑒𝑎𝑠𝑒_𝑐ℎ𝑎𝑛𝑔𝑒𝑖,𝑡−1 + 𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡−1 + 𝑌𝑒𝑎𝑟 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 +
𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝑒𝑖𝑡 (1)
The fitted value from the above regression is an estimate of the expected investments in leases in year t,
which we denote Lease_p, based on firm characteristics in year t-1. The regression residual is the
estimate of over- or under-investment in leases in year t, which we denote Lease_r. To capture growth
opportunities, we include V/P, a ratio of value to price.11 We follow Richardson (2006) by including
control variables that have been shown in prior research to shape investment decisions, including
leverage, the level of cash, firm age and size, and stock returns (Barro 1990; Bates, 2005; Hubbard, 1998;
Lamont, 2000). We augment the Richardson (2006) model with lagged changes in leases as a proxy for
lease renewals. We also include lagged sales growth as another proxy for growth opportunities (Biddle,
Hilary, and Verdi 2009). Industry, and year indicator variables capture additional variation in
investments in leases that are not explained by growth opportunities and financing constraints.12
In Table 2, Panel A provides descriptive statistics and Panel B provides correlations among the variables
we use to estimate the lease investment model. Panel C presents the results from estimating the model.
These results profile the type of firm that increased investments in leases during a period—namely,
11 Following Richardson (2006), V/P is the estimated value of equity (V) divided by market value of equity. We estimate V using the Ohlson (1995) framework, as (1–α r)BV+ α (1+r)X – α rd where, α equals(ω/(1+r–ω)); r equals 12%; and ω equals 0.62. ω is the abnormal earnings persistence parameter from the Ohlson (1995) framework. BV is the book value of common equity (CEQ), d is annual dividends (DVC) and X is operating income after depreciation (OIADP). 12 Our results remain robust when we also include firm fixed effects in estimating the Richardson (2006) model.
16
smaller and younger firms with less cash, stronger past sales growth and stock returns, and better
investment opportunities. The model has an adjusted R-square of 13.9 percent, suggesting it does a
reasonable job of capturing the determinants of firms’ future investments in leases.
For comparative analyses, we also examine over-investment through capital expenditures.
Specifically, we estimate Richardson’s overinvestment model:
𝑖_𝑛𝑒𝑤𝑖,𝑡 = 𝛼 + 𝛽1𝑉/𝑃𝑖,𝑡−1 + 𝛽2𝐿𝐸𝑉𝑖,𝑡−1 + 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1 + 𝛽4𝐴𝐺𝐸𝑖,𝑡−1 + 𝛽5𝑆𝐼𝑍𝐸𝑖,𝑡−1 + 𝛽6𝑅𝐸𝑇𝑈𝑅𝑁𝑖,𝑡−1 +
𝛽7𝑖_𝑛𝑒𝑤𝑖,𝑡−1 + 𝑌𝑒𝑎𝑟 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝑓𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝑒𝑖𝑡 (2)
The fitted value from the above regression is an estimate of the expected capital investments in year t,
which we denote ihat_p, based on firm characteristics in year t-1. The regression residual is the estimate
of over- or under-investment in year t, which we denote ihat_r.
4. Empirical Tests and Results
4.1 Cross-Sectional Tests of Future Performance
4.1.a Cross-Sectional Tests of Future Performance: One-Year Ahead Changes in Sales and Earnings
We begin our empirical tests by examining the relationship between over-investments in leased
assets and changes in the firm’s future performance. We estimate the following general model, varying
the dependent variable:
𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+1 = 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽3𝐼𝐵𝑖,𝑡 + 𝛽4𝑆𝑎𝑙𝑒𝑠𝑔𝑟𝑖,𝑡 + 𝛽5𝐴𝐶𝐶𝑖,𝑡 +
𝛽6𝑆𝑖𝑧𝑒𝑖,𝑡 + 𝑌𝑒𝑎𝑟 𝐹𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝐼𝑛𝑑𝑢𝑠𝑡𝑟𝑦 𝐹𝑖𝑥𝑒𝑑 𝑒𝑓𝑓𝑒𝑐𝑡𝑠 + 𝑒𝑖𝑡+1 (3)
The future performance metrics we examine include one-year-ahead changes in sales and net income
before extraordinary items, each measured as next year’s value less the current year’s value, deflated by
current year total assets.13 We measure the expected and residual investments in leases from equation
(1). We control for current earnings, sales growth, accruals, and size (similar to Doyle, Lundholm, and
13 Note that we compute the change before scaling by assets. Thus, the dependent variables cannot be interpreted as a change in asset turnover or a change ROA, and our results are not driven by a change in the asset base from one year to the next.
17
Soliman, 2003). If managers over-invest to enjoy the benefits that accompany firm size (Jensen 1986)
and if Lease_r captures over-investments in leased assets, then we predict that current lease
overinvestment will relate positively to future sales changes, but negatively to future earnings changes.
To further our understanding of the potential effects that over-investments in leases have on
performance separate from other forms of over-investment, we include the expected and residual
investments in property, plant, and equipment from equation (2) and include them in equation (3). We
estimate the following model:
𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+1 = 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒𝑝𝑖,𝑡+ 𝛽3𝑖ℎ𝑎𝑡𝑟𝑖,𝑡 + 𝛽4𝑖ℎ𝑎𝑡𝑝𝑖,𝑡
+ 𝛽5𝐼𝐵𝑖,𝑡
+𝛽6𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽7𝐴𝐶𝐶𝑖,𝑡 + 𝑒𝑖𝑡+1 (4)
This specification allows us to compare and distinguish the over-investment effect for leases to that of
capital expenditures.
Table 3, Panel A provides descriptive statistics and Panel B provides correlations. The mean
(median) lease obligation equals $394.7 million ($68.0 million). The typical firm experienced a change in
one-year-ahead sales averaging 8.8 percent of total assets, and a one-year-ahead change in earnings
averaging 1.1 percent of total assets. The correlations in Panel B provide univariate evidence to suggest
that over-investment in leases relate positively to one-year-ahead sales changes (0.151) and negatively
to changes in one-year-ahead income before extraordinary items (-0.061).
Panel C presents the results from estimating equation (3). We expect over-investment to be
concentrated in the lease changes unrelated to the firm’s renewals and investment opportunities
(estimated with Equation 1, in Table 2). In the first column of results, we find unexpected lease changes
relate to higher sales growth one year ahead (0.400, t-statistic=8.50). However, in the second column of
results, we find that unexpected changes in leases relate to lower earnings growth one year ahead (-
0.143, t-statistic=-5.37). Expected lease changes relate positively to future sales changes (0.683, t-
statistic=4.62) but have a neutral effect on future earnings changes (-0.143, t-statistic=-1.18).These
18
results are consistent with our predictions for over-investments in leases; the incremental sales growth
from additional leases do not compensate for the additional operating costs and lease expenses.
In columns (3) and (4), we examine the impact that unexpected lease changes have on future
performance after controlling for expected and unexpected CapEx (estimated with Equation 2). We find
similar results for unexpected lease changes - they relate to higher sales growth one year ahead (0.359,
t-statistic=7.64) and lower earnings growth one year ahead (-0.104, t-statistic=-4.01). In addition,
unexpected CapEx relates to higher sales growth one year ahead (0.136, t-statistic=5.42) and to lower
earnings growth one year ahead (-0.115, t-statistic=-6.41). These results suggest that over-investment in
leased assets is not simply a manifestation of over-investment in capital expenditures but a separate
phenomenon.
4.1.b Cross-Sectional Tests of Future Performance within Subsamples: Over- versus Under-investment
In this section, we examine over-investment in leased assets more deeply by splitting the
sample into two subsamples based on the sign of the unexpected lease change (Lease_r). We conduct
more powerful tests by examining whether the effects of over-investments are increasing within the set
of firms with positive unexpected increases in leases. These are more demanding tests because they
examine sales and earnings consequences that arise from over-investments in leases—not just whether
firms over-invest, but by how much firms over-invest.
We sort our subsample of firms with positive unexpected lease changes (Lease_r > 0) into
quintiles each year. For each quintile, we calculate the average change in sales and the average change
in earnings next year. We depict the results in Figure 1, and show that next-period sales changes are
increasing and next-period earnings changes are decreasing with the amount of over-investment in
leases. These patterns suggest an over-investment story—firms generate greater future sales increases
but lower future earnings changes with increasing amounts of over-investment in leased assets.
19
Our regression-based analyses with controls support these simple univariate patterns. Table 3,
Panel D reports the results from examining the relations between one-year-ahead changes in sales and
earnings and unexpected changes in leases within the two subsamples. Within the subsample of firms
that appear to over-invest in leases (Lease_r>0, in the first and fifth columns of results), future changes
in sales are increasing in unexpected lease changes (0.249; t-statistic = 3.67), and future changes in
earnings are decreasing (-0.165; t-statistic = -4.39). In addition, these results hold after controlling for
expected and unexpected CapEx (results in columns three and seven). These regression results suggest
that, among the firms that seemingly over-invest in leases, larger amounts of over-investment relate to
larger future sales increases, but lower future earnings changes.
We also examine the subsample of firms that under-invest in leased assets (Lease_r<0, columns
two and six of results). Keep in mind that for these tests, because the independent variables are
negative numbers (Lease_r<0), the results should be interpreted in the opposite directions. We find
that, by contrast, under-investments in leases relate negatively to future changes in sales (0.606; t-
statistic =4.31), and relate positively to future changes in earnings (-0.160; t-statistic = -2.15). However,
after controlling for expected and unexpected over-investment in CapEx (columns four and eight),
under-investments in leases are no longer significantly related to future changes in earnings.
Table 3, Panel D also reports the results from examining the relations between one-year-ahead
changes in sales and earnings and predicted changes in lease (Lease_p) within the two subsamples. In
both subsamples, future changes in sales are strongly increasing in predicted lease changes, but
predicted lease changes have a neutral relation with future changes in earnings.
4.1.c Cross-Sectional Tests of Longer-Run Future Performance
It may take more than one year for some leased assets to reach their full productive potential
for future sales and earnings growth. Periodic lease payments usually begin to accrue upon the
inception of the lease, but the leased asset may not generate sales for weeks or months, and may not
20
reach its full revenue and earnings generating potential until well after the first year. To capture the
impact of leases on longer-term firm performance we extend the analysis by examining the changes in
sales and earnings for years +2 and +3.
The results in Table 4 reveal that the inferences for year +1 are robust and persist to years +2
and +3 as well. Expected and unexpected lease changes are positively associated with changes in sales
during years +2 and +3. Unexpected lease changes are negatively associated with future earnings
changes in year +2 and year +3. Expected lease changes have a neutral association with changes in
earnings for year +2 or year +3. These results suggest that the sales growth and earnings consequences
of over-investments in leased assets in year t continue for at least three years after the investment, and
that the over-investment effects seem to be concentrated in unexpected lease changes.
The last four columns of Table 4 examine the subsample of observations with over-investments
in leased assets, focusing on positive unexpected lease changes. Within the subsample of unexpected
lease increases, we find positive but not significant associations with changes in sales in year +2 (t=1.54)
and year +3 sales (t=1.17). However, we do find significant negative relations with future earnings
changes in year +2 (t=-2.48) and year +3 (t=-1.80). Within this subsample, expected lease changes
continue to have a positive association with changes in sales through year +2 and year +3 sales, and are
not related to changes in earnings in year +2 or year +3. Overall, these results indicate that unexpected
investments in leases that are unrelated to the firm’s investment opportunities are associated with
declining future earnings over a three-year horizon, consistent with over-investment.
4.2 Cross-Sectional Contemporaneous Returns
In our second set of tests, we examine the extent to which the market incorporates the change
in leased assets and obligations into share price. If the market observes the firm increasing its
investments in leases and understands the negative implications for future earnings, we expect to
observe a negative relation with contemporaneous returns. However, increases in lease obligations also
21
increase the cost of capital. Prior research has established that an increase in the present value of lease
obligations is associated with increased leverage and credit risk, as well as higher costs of equity capital
(Bratten et al 2013; Dhaliwal et al 2011), which should trigger a decline in share price. A negative
relation between lease changes and returns could therefore reflect greater leverage (a discount rate
effect), lower expected future cash flows because of lower future earnings (a cash flow effect), or both.
4.2.a Controlling for Changes in Expected Returns
Our prior results show over-investments in leased assets have a strong negative effect on future
earnings, and therefore on future cash flows. We seek to isolate and test whether the capital markets
detect and price the negative cash flow effects from over-investments in leases, apart from the discount
rate effects. Therefore, we need to separate the discount rate and cash flow effects of over-
investments in leases on stock returns.
Isolating the effects of changes in lease obligations on discount rates from the effects on
expectations of future cash flows poses an interesting challenge. To do this, we rely on a proxy for the
change in the market’s expected return, Expret_chg, developed in Penman and Yehuda (2016). Penman
and Yehuda (2016) estimate expected returns using the following cross‐sectional regression model:
𝑅𝑡 = 𝛼 + 𝛽1𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠𝑡−1
𝑃𝑡−1+ 𝛽2
𝐵𝑡−1
𝑃𝑡−1+ 𝛽3𝐴𝐶𝐶𝑅𝑡−1 + 𝛽4𝐼𝑁𝑉𝐸𝑆𝑇𝑡−1 + 𝛽5∆𝑁𝑂𝐴𝑡−1 + 휀𝑡 (5)
Following Penman and Yehuda (2016), Rt equals the raw buy-and-hold return for year t,
calculated as compounded monthly returns from CRSP over the period from three months after the
beginning of the fiscal year t to three months after the end. Earningst-1 equals earnings before
extraordinary items (Compustat item IB) and special items (item SPI), minus preferred dividends (item
DVP), with a tax allocation to special items at the prevailing federal statutory corporate income tax rate
for the year. Earningst-1/Pt-1 is the earnings yield for fiscal year t. B t-1/P t-1 equals book value of common
22
equity divided by market value of equity. ACCR t-1 equals accruals divided by average total assets.14
INVESTt-1 equals investment calculated as the change in gross property, plant, and equipment (item
PPENT) plus the change in inventory (item INVT)) divided by lagged assets. ΔNOA t-1 equals the change in
net operating assets divided by average total assets. Although we follow Penman and Yehuda’s (2016)
methodology faithfully, we make two adaptations to incorporate the fact that leased assets are
operating assets. Specifically, we include the Lease_change value in both the investment measure
(INVEST) and the change in net operating assets measure (NOA).
Following Penman and Yehuda (2016), we estimate this cross-sectional returns model annually,
and compute the mean coefficients over a rolling 10‐year period prior to year t-1. We then estimate the
expected return for each firm in year t-1 by applying the estimated coefficients in regression (5) to
observed accounting variables for each firm out of sample:
𝐸𝑡−1(𝑅𝑡) = �̂� + 𝛽1̂𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠𝑡−1
𝑃𝑡−1+ 𝛽2̂
𝐵𝑡−1
𝑃𝑡−1+ 𝛽3̂𝐴𝐶𝐶𝑅𝑡−1 + 𝛽4̂𝐼𝑁𝑉𝐸𝑆𝑇𝑡−1 + 𝛽5̂∆𝑁𝑂𝐴𝑡−1 (6)
We then estimate the expected return for year t the same way, using the rolling mean coefficients
estimated over the 10 years prior to year t. We then compute the change in expected return during
period t as ΔEt(Rt+1)= Et(Rt+1) - Et-1(Rt). We use this measure as the control for changes in firms’ discount
rates, in order to isolate the cash flow effects of over-investments in leases on stock returns.
In Table 5, Panel A provides descriptive statistics of the variables we use to estimate the
expected returns model, and Table B provides correlations.15 Each of the variables exhibit the expected
correlation in Panel B, with the exception of the relationship between Hret_1 and Earningst/Pt in the
Pearson correlation. In Panel C, we report the model estimation results we use to compute the change
in expected returns, including the mean coefficient estimates and t-statistics. For their 1963-2012
14 We measure accruals as the sum of the change in accounts receivable (item RECT), the change in inventory (item INVT), and the change in other current assets (item ACO), minus the sum of the change in accounts payable (item AP) and the change in other current liabilities (item LCO), minus depreciation and amortization expense (item DP). 15 The sample we use to estimate the expected returns model consists of all non-financial firms with market value of equity greater than $50 million, price greater than $1, positive shareholders’ equity, and the model variables.
23
sample period, Penman and Yehuda (2016) report a positive relation between current returns and the
lagged earnings yield and the lagged book to price ratio, and a negative relation with lagged accruals,
lagged investment, and lagged change in net operating assets. Penman and Yehuda (2016) report an
adjusted R-square of 4 percent. For our 2001 to 2014 sample period, we find also find a marginal
positive relation for the lagged earnings yield and a somewhat stronger positive relation for the lagged
book to price ratio. In contrast, the accruals, investment, and change in net operating assets variables
are not significant. Like Penman and Yehuda (2016), we report an adjusted R-square of 4.4 percent.
4.2.b Testing the Relation between Investments in Leases and Stock Returns
We next turn to our tests of the association between over-investments in leases and stock
returns. To control for the discount rate effects, we incorporate the change in expected returns
(Expret_chg), as described above. We also control for various other known determinants of stock
returns, including size (MVE) and the book-to-market ratio (BTM), as well as all of the control variables
and fixed effects in our tests of future sales and earnings growth. We estimate the following model:
𝐶𝑆𝐴𝑅𝑖,𝑡 = 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝑖ℎ𝑎𝑡_𝑟𝑖,𝑡 + 𝛽4𝑖ℎ𝑎𝑡_𝑝𝑖,𝑡 + 𝛽5𝐸𝑥𝑝𝑟𝑒𝑡_𝑐ℎ𝑔𝑖,𝑡
+𝛽6𝐼𝐵𝑖,𝑡 + 𝛽7𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽8𝐴𝐶𝐶𝑖,𝑡 + 𝛽9𝑟𝐵𝑇𝑀𝑖,𝑡−1 + 𝛽10𝑟𝑀𝑉𝐸𝑖,𝑡−1 + 𝑒𝑖𝑡 (7)
We calculate abnormal returns, CSAR, for each firm-year by compounding the firm’s raw return
over a one-year holding period and then subtracting the return on the CRSP size-based decile portfolio
to which the firm belongs at the beginning of the holding period.16 We begin cumulating returns on the
first day of the fourth month after the prior fiscal year-end. This allows the market time to impound the
financial statement information into share prices. We end the accumulation period one year later and
thereby capture price movements caused by investors reacting to information and events during the
year that confirm or refute their beliefs about the firm’s value. If a firm delists, we include any de-listing
16 Berk (1995) suggests that firm size is a catch-all risk proxy.
24
return in the calculation of returns, place the funds available after delisting into the size-based decile,
and continue cumulating returns through the period.
We expect the coefficient on Lease_r will relate negatively to stock returns. This would suggest
the market incorporates the negative cash flow news from over-investments in leases into share prices,
after controlling for discount rate news.
In Table 5, Panel D, we report results from examining the relation between lease changes and
contemporaneous returns. In the first column of results, we document an insignificant negative relation
between unexpected lease changes and returns (-0.078, t-statistic=-0.62) and a negative relation with
expected lease changes (-2.525, t-statistic= -6.02).
To control for changes in discount rates, we include our estimate of Penman and Yehuda’s
(2016) change in expected return (Expret_chg) in column 2. This variable should capture the relation
between stock returns and changes in firms’ expected returns (discount rate news), allowing the lease
change variables to capture the relation between stock returns and changes in the market’s
expectations about future performance associated with lease changes (cash flow news). Interestingly,
the returns results in column 2 reveal that capital market prices impound very different implications in
expected versus unexpected lease changes. The relation between stock returns and unexpected lease
changes becomes significantly negative (-0.420, t-statistic=-3.26) while the relation with expected lease
changes becomes negative but not significant (-0.671, t-statistic=-1.50). These results suggest the
market prices over-investments in leases as value-destructive, consistent with our prior results on future
earnings performance. Column three reports similar results after controlling for the impact of expected
and unexpected investment in PPE.
In column four, we examine the relation between unexpected lease changes and
contemporaneous returns within the subsample of firms that over-invest in leases (Lease_r>0). We
observe a significant negative relation between unexpected leases changes and contemporaneous
25
returns (-0.535, t-statistic=-2.31). Overall, these results indicate the capital markets detect and price the
negative cash flow effects from over-investments in leases, apart from the discount rate effects.
4.3 Executive Compensation
In tables 3 and 4, we showed that over-investments in leases produce higher sales growth, but
lower earnings growth, from one-year ahead through three years ahead. In this section, we examine the
link between sales and executive compensation for our sample firms. As we discuss in section 2,
executive compensation may provide one potential motive (among many) for over-investment in leased
assets. Because sales growth commonly plays a prominent role in elements of executive pay, some
managers have incentives to over-invest in leased assets to drive sales growth and increase their pay.
To examine the relation between sales and CEO compensation for our sample firms, we adapt
the Core, Holthausen, and Larcker (1999) model. Specifically, we estimate the following models:
𝐶𝑂𝑀𝑃𝑖,𝑡+1 = 𝛼 + 𝛽1𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝑅𝑂𝐴𝑖,𝑡 + 𝛽3𝑀𝑇𝐵𝑖,𝑡 + 𝛽4𝑅𝐸𝑇𝑖,𝑡 + 𝛽5𝑅𝑂𝐴_𝑣𝑜𝑙𝑖,𝑡 + 𝛽6𝑅𝐸𝑇_𝑣𝑜𝑙𝑖,𝑡−1 +
𝑒𝑖,𝑡+1 (8)
𝐶𝑂𝑀𝑃𝑖,𝑡+1 = 𝛼 + 𝛽1𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝐷𝑅𝐸𝑆𝑖,𝑡 + 𝛽3𝐷𝑅𝐸𝑆_𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽4𝑅𝑂𝐴𝑖,𝑡 + 𝛽5𝑀𝑇𝐵𝑖,𝑡 + 𝛽6𝑅𝐸𝑇𝑖,𝑡 +
𝛽7𝑅𝑂𝐴_𝑣𝑜𝑙𝑖,𝑡 + 𝛽8𝑅𝐸𝑇_𝑣𝑜𝑙𝑖,𝑡−1 + 𝑒𝑖𝑡+1 (9)
and
𝐶𝑂𝑀𝑃𝑖,𝑡+1 = 𝛼 + 𝛽1𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝐷𝑅𝐸𝑆𝑖,𝑡 + 𝛽3𝐷𝑅𝐸𝑆_𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝐷𝑖ℎ𝑎𝑡𝑖,𝑡 + 𝛽3𝐷𝑖ℎ𝑎𝑡_𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 +
𝛽4𝑅𝑂𝐴𝑖,𝑡 + 𝛽5𝑀𝑇𝐵𝑖,𝑡 + 𝛽6𝑅𝐸𝑇𝑖,𝑡 + 𝛽7𝑅𝑂𝐴_𝑣𝑜𝑙𝑖,𝑡 + 𝛽8𝑅𝐸𝑇_𝑣𝑜𝑙𝑖,𝑡−1 + 𝑒𝑖𝑡+1 (10)
COMP equals one-year-ahead total compensation as reported in SEC filings, which equals the sum of
SALARY, BONUS, STOCK_AWARDS, OPTION AWARDS, NONEQ_INCENT, PENSION_CHG, and OTHCOMP.
Sales is revenues (in $ millions) for the year prior to the year in which compensation is awarded. DRES
equals 1 if Lease_r is greater than zero, and 0 otherwise; DRES_Sales is an interaction term equaling
Sales multiplied by DRES. Dihat equals 1 if ihat_r is greater than zero, and 0 otherwise; Dihat_Sales is an
26
interaction term equaling Sales multiplied by Dihat. MTB equals the market to book ratio averaged over
the five years ended the year prior to the year of the CEO’s compensation. ROA equals the ratio of
income before extraordinary items to average total assets for the prior year. RET is the stock return for
the prior year. ROA_vol is the standard deviation of ROA for the prior five years. RET_vol is the standard
deviation of annual stock market return for the prior five years.
We collect compensation data for our sample firms from ExecuComp. ExecuComp only includes
S&P 1500 firms, so our sample for this analysis only includes 3,456 firm-year observations.17 Table 6
Panel A reports descriptive statistics for our regression variables. The average total CEO compensation
for our sample firms was $4.72 million. The typical firm in this sample generated $5,327 million in sales,
had a five-year moving average market-to-book ratio of 3.3, and generated an ROA of 5.5 percent.
We report the results in Table 6 Panel B. The first column examines the role of sales in CEO
compensation. Consistent with prior literature, sales has a very strong positive relation with
compensation in the subsequent year (0.140; t-statistic = 23.92). The second column of Table 6 includes
an indicator (DRES) for observations with positive residual lease changes, our proxy for over-investments
in leases. These firms do not exhibit higher levels of executive compensation, as the DRES main effect
equals 10.055 (t-statistic = 0.06). However, executive pay for these firms has an elevated sensitivity to
sales, as the DRES*Sales interaction is significantly positive (0.066; t-statistic = 5.43). The combined
effect of sales on subsequent compensation is strongly positive. In column three, we control for firms
that over-invest in CapEx. The DRES interaction remains significantly positive (0.077; t-statistic = 6.23).
To corroborate these findings, in the final column of Panel B we report results from estimating (8) with
the set of observations with positive residual lease changes. Within this subsample, Sales are a strong
leading indicator of CEO compensation in the subsequent year. Among these firms, Sales has a positive
17 In untabulated analyses, we confirm that our main results in Table 3, showing future sales growth is increasing but future earnings growth is decreasing in over-investments in leased assets, hold within this subsample.
27
relation with Comp of 0.188 (t-statistic = 17.51). Executives of firms that over-invest in leased assets
derive direct benefits from higher sales in the form of higher compensation.
4.4 Additional Tests and Robustness
4.4.1 Free Cash Flows and Overinvestment
Table 7 examines how overinvestments in leases relate to free cash flows and overinvestments
in CapEx. In Panel A of Table 7, we first sort firms into quintiles based on lease_r. If leasing and capital
expenditures are competing avenues of over-investment, we expect to observe high (low) levels of
ihat_r in low (high) lease_r quintiles. Alternatively, if firms pursue over-investment strategies through
leasing primarily when free cash flows are unavailable for CapEx, we expect free cash flows to be low in
the high lease_r quintile. In contrast, we find that the low (high) lease_r quintile has the lowest (highest)
values of ihat_r and free cash flows. This suggests over-investment in leased assets coincides with over-
investment in CapEx and with strong free cash flows. This could reflect a need for free cash flows and
CapEx for leasehold improvements to prepare leased assets for their intended use.
The previous results suggest that over-investment in leased assets and over-investment in CapEx
may share a similar sensitivity to free cash flows. However, a sort of firms into quintiles based on free
cash flows reveals that this is not the case. As shown in Figure 2, although both unexpected lease
changes (lease_r) and unexpected CapEx (ihat_r) increase across free cash flow quintiles, the slope is
steeper for residual CapEx. Firms with low free cash flows are much more likely to under-invest in CapEx
whereas firms with high free cash flows are much more likely to over-invest. In particular, as Panel B of
Table 7 reports, comparing lease_r across the firms with lowest quintile of free cash flows versus the
highest quintile firms, we observe a difference of 0.012 (p < 0.001; a difference of 1.2% of total assets).
In contrast, comparing ihat_r across extreme quintiles of free cash flows reveals a difference of 0.036 (p
< 0.001; 3.6% of total assets), an effect three times the magnitude of lease_r.
28
In Panel C of Table 7, we regress over-investment in leases and capital expenditures on free cash
flows, allowing the coefficient to vary based on the sign of free cash flows. Jensen’s (1986) free cash
flow hypothesis predicts that firms will over-invest in capital expenditures when free cash flows are
positive. Thus, the sensitivity of capital expenditures should be asymmetric and more pronounced when
free cash flows are positive. In contrast, over-investment through leasing uses relatively less cash; when
a firm wishes to deploy high levels of cash, leasing is an unlikely avenue. The results in Panel C are
consistent with this prediction. Specifically, although lease_r is sensitive to both positive and negative
free cash flows, no evidence of asymmetry exists. In contrast, ihat_r is highly sensitive to free cash flows
when free cash flows are positive.
4.4.2 Future Returns
Given that we show that over-investments in leases can negatively impact future earnings for at
least three years, it begs a question: How long does it take the market to fully price the implications of
over-investments in leases for future performance? Ge (2006) examines the equity pricing implications
of operating leases and finds that larger increases in operating lease obligations portend lower future
stock returns. Although the evidence on debt ratings, cost of debt, and cost of equity suggest that
capital markets price disclosed information about leases, Ge’s (2006) results suggest that this
information is not fully priced in equity markets. In an untabulated analysis, we examine one-year-ahead
returns. We find unexpected lease changes are not related to future returns (0.030, t-statistic=0.26,
untabulated). We find similar results (0.115, t-statistic=0.98, untabulated) when we control for the
change in expected return based on the Penman and Yehuda (2016) approach. Thus, our evidence of a
negative relation between over-investments in leases and contemporaneous returns does not continue
to manifest in returns one year ahead.
4.4.3 Discount rates
29
In our main tests, we compute the present value of the operating lease commitments using
time-invariant firm-specific discount rates calculated as the median interest rate that the firm
experienced during the sample period. To ensure that this research design choice does not drive our
results, we re-estimate our tests under several alternative discount rate choices. We allow for a firm-
and time-specific discount rate, constant discount rates from 6 percent to 10 percent across all firms
(similar to prior lease studies, such as Bratten et al. 2013), and discount rate of 0 percent (i.e.,
undiscounted sum of expected future lease payments. Our inferences do not change across these
various iterations. Results are available upon request.
4.4.4 Five percent of total asset threshold
In our main results, we eliminate a firm if leasing is not an important part of its business: the
undiscounted sum of expected lease payments is less than five percent of total assets. To test whether
our results are robust to this sample selection criterion, we replaced our five percent cutoff with a four
percent cutoff, a three percent cutoff, and no cutoff (zero percent). Our inferences (untabulated) are
robust to these alternative cutoff assumptions.
5. Conclusion
Building on academic literature from several areas (including executive compensation, agency,
asset pricing and leases), we ask: Do firm managers use operating leases to “over-invest” in assets to
drive revenue growth? If so, what are the consequences, if any?
In our empirical analyses, we develop and estimate a model to predict the expected amount of
investment in leased assets, given the firm’s economic and strategic characteristics. We use our model
to parse new investments in leased assets into expected and unexpected leases. If our model for
expected leases is descriptive, positive residuals should estimate over-investments in leased assets. We
predict and find that our proxy for over-investments in leased assets is strongly associated with future
30
sales growth, and future increases in CEO compensation, even though it is negatively associated with
future earnings growth and contemporaneous stock returns.
We corroborate our main findings with various additional tests. For example, we predict and
find that unexpected lease changes are associated with lower future earnings changes over the
following three years. Also, in our tests of contemporaneous returns, we control for current period
changes in expected returns (the discount rate effect), to isolate and test the market’s pricing of the
(negative) cash flow effects from over-investments in leases. Controlling for changes in discount rates of
leasing firms, the value destruction effects from over-investments in leases become even stronger.
Overall, our evidence should interest investors, boards and managers of firms engaged in
leasing, participants in the institutions that govern such firms, and researchers. For investors, we show
that lease changes can generate positive future sales changes but negative earnings changes. Further,
we also develop an approach to distinguish leases related to investment opportunities from leases
reflecting suboptimal investment. From a corporate governance perspective, our results should increase
awareness of the potential for over-investment among companies engaged in leasing. For regulators
and standard setters, who have been facing much criticism over accounting recognition of leases, our
results provide an additional justification for the new standards that will require recognition of assets
and obligations under most operating leases. Prior research shows that improved financial reporting
quality reduces suboptimal investment (Biddle and Hilary 2006; Biddle, Hilary, and Verdi 2009).
Improving financial reporting through on-balance sheet recognition of lease rights and obligations could
potentially curtail the over-investment effects we document. For researchers, we add novel evidence on
the economic consequences of leasing. Additionally, we contribute to the research on potential agency
problems by showing that the over-investment problem can extend beyond investment of free cash
flows.
31
References
Altamuro, J., R. Johnston, S. Pandit, and H. Zhang. 2014. Operating Leases and Credit Assessments.
Contemporary Accounting Research 31: 551-580
Baker, G., Jensen, M., & Murphy, K. (1988). Compensation and Incentives: Practice vs. Theory. The
Journal of Finance, 43(3), 593-616.
Barro, R. 1990. The stock market and investment. Review of Financial Studies 3:115-130.
Barth, M. E., W. H. Beaver, and W. R. Landsman. 1998. Relative valuation roles of equity book value and
net income as a function of financial health. Journal of Accounting and Economics 25: 1–34.
Bates, T. 2005. Asset sales, investment opportunities, and the use of proceeds. Journal of Finance
60:105-135.
Baumol, W. 1959. Business Behavior, Value, and Growth (Macmillan, New York).
Berk, J. 1995. A Critique of Size-Related Anomalies, The Review of Financial Studies, Volume 8, Issue 2:
275–286
Berle, A. and G. Means. 1932. The Modern Corporation and Private Property. (MacMillan, New York).
Biddle, G. C., G. Hilary and R.S. Verdi. 2009. How Does Financial Reporting Quality Relate to Investment
Efficiency? Journal of Accounting and Economics 48:112-131.
Biddle, G.C., and G. Hilary. 2006. Accounting-Quality and Firm-Level Capital Investment. The Accounting
Review 81(5):963-982.
Blanchard, O.J., F. Lopez-de-Silanes and A. Shleifer. 1994. What do firms do with cash windfalls?
Journal of Financial Economics 36:337−360.
Bond, S., and C. Meghir. 1994. Dynamic investment models and the firm’s financial policy. Review
of Economic Studies 61:197−222.
Bowman, R. 1980. The debt equivalence of leases: An empirical investigation. The Accounting Review 55
(2): 237-253.
Bratten, B., P. Choudhary, and K. Schipper. 2013. Evidence that Market Participants Assess Recognized
and Disclosed Items Similarly when Reliability is Not an Issue. The Accounting Review 88: 1179-
1210.
Callahan, C. M., R. E. Smith, and A. W. Spencer. 2013. The valuation and disclosure implications of FIN46
for synthetic lease liabilities. Journal of Accounting and Public Policy 32 (4): 271–291.
Calomiris, C. and R. Hubbard. 1995. Internal finance and investment: evidence from the undistributed
profits tax of 1936-37. Journal of Business 68: 443-482.
Campbell, J., and R. Shiller. 1988. Stock Prices, Earnings, and Expected Dividends. Journal of Finance 43,
661-676.
Campbell, J, and T. Vuolteenaho. 2004. Bad Beta, Good Beta. American Economic Review 94 (5), 1249-
1275.
Carter, M., L. Lynch, and I. Tuna. 2007. The Role of Accounting in the Design of CEO Compensation. The
Accounting Review 82(2):327-357.
Chen, C., D. Young and Z. Zuang. 2013. Externalities of Mandatory IFRS Adoption: Evidence from Cross-
Border Spillover Effects of Financial Information on Investment Efficiency. Accounting Review 88:
881-914.
Chen, F., O.-K. Hope, Q. LI and X. Wang. 2011. Financial Reporting Quality and Investment Efficiency of
Private Firms in Emerging Markets. Accounting Review 86: 1255-1288.
32
Chen, Wei, P. Hribar, and S. Melessa. Forthcoming. Incorrect Inferences When Using Residuals as
Dependent Variables. Journal of Accounting Research.
Chirinko, R. 1995. Why does liquidity matter in investment equations? Journal of Money, Credit, and
Banking 27: 527-548.
Chu, L., T. J. Levesque, R. Mathieu, and P. Zhang. 2008. The Impact of Lease Obligations on Loan Pricing.
Working paper, Wilfrid Laurier University and University of Toronto.
Core, J., Holthausen, R. and Larcker, D. (1999). Corporate governance, CEO compensation and firm
performance. Journal of Financial Economics, 51, 371-406.
Dhaliwal, D., H.S. Lee, and M. Neamitu. 2011. The Impact of Operating Leases on Firm Financial and
Operating Risk. Journal of Accounting, Auditing, and Finance 26:151-197.
Donaldson, G. 1984. Managing Corporate Wealth (Praeger, New York).
Doyle, J.T., Lundholm, R.J. & Soliman, M.T. Review of Accounting Studies. 2003 8: 145-174.
Elam, R. 1975. The effect of lease data on the predictive ability of financial ratios. The Accounting Review
50 (1): 25–42.
Ely, K. M. 1995. Operating lease accounting and the market’s assessment of equity risk. Journal of
Accounting Research 33 (2): 397–415.
Financial Accounting Standards Board. 2016. ASU No. 2016-02. Leases (Topic 842). Norwalk, CT:FASB.
Ge, Weili, Activities, Earnings Persistence and Stock Prices: Evidence from Operating Leases (November
22, 2006). Working paper. University of Washington.
Gilchrist, S., and C.P. Himmelberg. 1995. Evidence on the role of cash flow for investment. Journal
of Monetary Economics 36:531−572.
Harford, J. 1999. Corporate cash reserves and acquisitions. The Journal of Finance 54 (6): 1969-1997.
Harris, M., and A. Raviv. 1990. Capital structure and the informational role of debt. Journal of
Finance 45:321−350.
Hart, O., and J. Moore. 1995. Debt and seniority: an analysis of the role of hard claims in constraining
Management. American Economic Review 85:567−585.
Huang, R., C. Marquardt, and B. Zhang. 2015 Using Sales Revenue as a Performance Measure (July 28,
2015). Working paper. Available at SSRN: https://ssrn.com/abstract=2636950
Hubbard, R. and D. Palia. 1995. Executive Pay and Performance: Evidence from the US
Banking Industry. Journal of Financial Economics 39(1) 105-30.
Imhoff, E., C. Lipe, and D. Wright. 1993. The effects of recognition versus disclosure on shareholder risk
and executive compensation. Journal of Accounting, Auditing, & Finance 8 (4): 335-368.
Kraft, P. 2015. Rating agency adjustments to GAAP financial statements and their effect on ratings and
credit spreads. The Accounting Review 90(2): 641–674.
Krishnan, V.S., Moyer, R.C. 1994.Bankruptcy Costs and the Financial Leasing Decision. Financial
Management. 23 (2):31–42
Jensen, M. 1986. Agency costs and free cash flow, corporate finance and takeovers. American Economic
Review 76: 659-665.
Jensen, M.C. 1993. The modern industrial revolution, exit, and the failure of internal control systems.
Journal of Finance 48:831−880.
Jensen, M.C., and W.H. Meckling. 1976. Theory of the firm: managerial behavior, agency costs and
ownership structure. Journal of Financial Economics 3:305−360.
33
Lamont, O. 2000. Investment plans and stock returns. Journal of Finance 55: 2719-2748.
Lang, L.H.P., R.M. Stulz and R.A. Walkling. 1991. A test of the free cash flow hypothesis: the case
of bidder returns. Journal of Financial Economics 29:315−335.
Lang, L., E. Ofek, and R. Stulz. 1996. Leverage, investment, and firm growth. Journal of Financial
Economics 40: 3-30.
Lewis, C. M., and J. S. Schallheim. 1992. Are debt and leases substitutes? Journal of Financial and
Quantitative Analysis 27 (4): 497–511.
Lewellen, W.G., C. Loderer and A. Rosenfeld. 1985. Merger decisions and executive stock ownership
in acquiring firms. Journal of Accounting and Economics 7:209−231.
Lipe, R. C. 1986. The information contained in the components of earnings. Journal of Accounting
Research 24 (3): 37–64.
Lipe, R. 2001. Lease accounting research and the G4+1 Proposal. Accounting Horizons 15 (3): 299-310.
Marris, R. 1964. The Economic Theory of Managerial Capitalism (Free Press, Glencoe, IL).
Marston, F., and R. S. Harris. 1988. Substitutability of leases and debt in corporate capital structures.
Journal of Accounting, Auditing & Finance 3 (2): 147–164.
Modigliani, F., and M. Miller. 1958. The cost of capital, corporation finance, and the theory of
Investment. American Economic Review 48:261−297.
Murphy, K. 1985. Corporate performance and managerial remuneration: An empirical analysis. Journal
of accounting and economics, 7(1-3), pp.11-42.
Murphy, K. 1998. Executive Compensation. Handbook of Labor Economics, Vol. 3, Orley Ashenfelter and
David Cards, editors (North Holland)
Myers, S.C., and N.C. Majluf. 1984. Corporate financing and investment decisions when firms have
information that investors do not have. Journal of Financial Economics 13:187−222.
Newey, W. K., and K. D. West. 1987. A simple, positive semi-definite,heteroskedasticity and
autocorrelation consistent covariance matrix. Econometrica 55: 703–708.
Ohlson, J. 1995. Earnings, book values and dividends in equity valuation .Contemporary Accounting
Research. 11:661-687
Opler, T., Pinkowitz, L., Stulz, R., and Williamson, R. 1999. The determinants and implications of
corporate cash holdings. Journal of Financial Economics 52 (1): 3-46.
Opler, T., Pinkowitz, L., Stulz, R., and Williamson, R. 2001. Corporate cash holdings. Journal of Applied
Corporate Finance 14 (1): 55-67.
Penman, S.H. and N. Yehuda. 2016. A Matter of Principle: Accounting Reports Convey Both Cash-Flow
News and Discount-Rate News. Working paper. Columbia University. (June).
Rapoport, M. 2013. Lease accounting may shift—Companies might have to increase amount of debt
they report under change. Wall Street Journal (May 17): C2.
Realdon, M. 2006. Pricing the credit risk of secured debt and financial leasing. Journal of Business
Finance & Accounting 33 (7/8): 1298–1320.
Richardson, S. 2006. Over-investment of Free Cash Flow. Review of Accounting Studies 11:159-189.
Schaller, H. 1993. Asymmetric information, liquidity constraints and Canadian investment. Canadian
Journal of Economics 26: 552-574.
Schallheim, J., K. Wells, and R. Whitby. 2013. Do leases expand debt capacity? Journal of Corporate
Finance 23 (December): 368–381.
34
Sengupta, P. and Wang, Z. 2011. Pricing of Off‐Balance Sheet Debt: How Do Bond Market Participants
Use the Footnote Disclosures on Operating Leases and Postretirement Benefit Plans? Accounting
& Finance Vol. 51, Issue 3, 787-808.
Sharpe, S. and Nguyen, H. 1995. Capital market imperfections and the incentive to lease. Journal of
Financial Economics 39:271-294
Sherman, B. and P. King. 2014. A Sure-Fire Way to Harm the Economy: A Proposal to Count Leases as
Liabilities. Wall Street Journal (November 14).
Skantz, T. R. 2012. CEO Pay, Managerial Power, and SFAS 123(R). The Accounting Review 87:2152-2179.
Spencer, A. and T. Webb. 2015. Leases: A Review of Contemporary Academic Literature Relating to
Lessees. Accounting Horizons 29:997-1023.
Stein, J. 2003. Agency, information, and corporate investment. Handbook of the Economics of Finance,
Edited by G. Constantinides, M. Harris, and R. Stulz (North Holland).
Stulz, R. 1990. Managerial discretion and optimal financing policies. Journal of Financial Economics 26: 3-
27.
United States Chamber of Commerce. 2012. The Economic Impact of the Current IASB and FASB
Exposure Draft on Leases.
Vuolteenaho, T. 2001. What drives firm-level stock returns. Journal of Finance 57, 233-264.
Williamson, O.E. 1964. The Economics of Discretionary Behavior: Managerial Objectives in a Theory
of the Firm (Prentice Hall, Englewood Cliffs, NJ).
Yan, A. 2006. Leasing and debt financing: Substitutes for or complements? Journal of Financial and
Quantitative Analysis 41 (3): 709–731.
Zwiebel, J. 1996. Dynamic capital structure under management entrenchment. American Economic
Review 86:1197−1215.
35
Appendix A
Example: Lease obligation calculation.
UNITED STATES
SECURITIES AND EXCHANGE COMMISSION
Washington, D.C. 20549
FORM 10-K
For the fiscal year ended December 31, 2012
SKYWEST, INC.
From: CONSOLIDATED BALANCE SHEETS (Dollars in thousands)
2012 2011
LONG TERM DEBT, net of current maturities 1,470,568 1,606,993
From: CONSOLIDATED STATEMENTS OF COMPREHENSIVE INCOME (LOSS)
Interest expense $77,380
From: Footnote 6, Commitments and Contingencies
Year ending December 31,
2013 387,999
2014 360,797
2015 309,378
2016 239,989
2017 182,291
Thereafter 720,733
2,201,187
From Item 1A. RISK FACTORS
We have a significant amount of contractual obligations.
As of December 31, 2012, we had a total of approximately $1.6 billion in total long-term debt obligations.
Substantially all of this long-term debt was incurred in connection with the acquisition of aircraft, engines and
related spare parts. We also have significant long-term lease obligations primarily relating to our aircraft fleet. These
leases are classified as operating leases and therefore are not reflected as liabilities in our consolidated balance
sheets. At December 31, 2012, we had 568 aircraft under lease, with remaining terms ranging from one to 13 years.
Future minimum lease payments due under all long-term operating leases were approximately $2.2 billion at
December 31, 2012. At a 4.7% discount factor, the present value of these lease obligations was equal to
approximately $1.8 billion at December 31, 2012. Our high level of fixed obligations could impact our ability to
obtain additional financing to support additional expansion plans or divert cash flows from operations and expansion
plans to service the fixed obligations. (emphasis added)
36
Appendix A (continued)
Example. Lease obligation calculation.
Yearly Thereafter Lease Obligation Calculation
Number of years past 2017 = Thereafter Amount divided by 2017 amount
= 720,733 / 182,291 = 3.95
We spread the 720,733 over the next 4 years in the following manner:
2018 2019 2020 2021 Sum of Years 2018 - 2021
182.3 182.3 182.3 173.9 720.733
Firm-Specific Discount Rate Calculation:
Yearly Interest Rate = xintt/((dlttt+dlttt-1)/2)
2012 Interest Rate: 77,380 / ((1,470,568 + 1,606,993) / 2) = 0.0503
Year: 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
Rate: 0.0495 0.0640 0.0891 0.0509 0.0467 0.0586 0.0758 0.0741 0.0629 0.0498 0.0487 0.0480 0.0503 0.0506
Median rate: 0.0507 (average of 2003 and 2013 interest rates)
Present Value Calculation:
2013 2014 2015 2016 2017 2018 2019 2020 2021
Lease Payment 387.999 360.797 309.378 239.989 182.291 182.291 182.291 182.291 173.86
PV Factor 0.926 0.857 0.794 0.735 0.681 0.630 0.583 0.540 0.500
Present Value 359.258 309.325 245.594 176.399 124.064 114.874 106.365 98.486 86.973
PV Lease Obligation 1,800.39
37
In this figure, we group firms into quintiles by positive unexpected lease changes. We adapt the Richardson (2006)
investment model to our sample to estimate unexpected lease changes. For each quintile, we plot the average
change in sales and the average change in earnings for the subsequent year. N=4,575 firm-years.
-0.05
0
0.05
0.1
0.15
0.2
0.25
0 1 2 3 4
Ch
ange
in N
ext-
Year
Sal
es o
r Ea
rnin
gs
(div
ided
by
curr
ent
year
ass
ets)
Positive Unexpected Lease Change Quintile
Figure 1. Next-year Changes in Sales and Earnings by Positve Unexpected Lease Change Quintile
Next-Year Sales Change Next-Year Earnings Change
38
In this figure, we group firms into quintiles by free cash flows. For each quintile, we plot the average lease residual
and the average Ihat residual. N=11,434 firm-years.
-0.025
-0.020
-0.015
-0.010
-0.005
0.000
0.005
0.010
0.015
0.020
Q1 Q2 Q3 Q4 Q5
Leas
e R
esid
ual
an
d Ih
at R
esid
ual
Free Cash Flow Quintile
Figure 2. Lease Residuals and Ihat Residuals by Free Cash Flow Quintiles
Lease_r Ihat_r
39
Table 1
Descriptive Statistics: Lease Obligations
Panel A: Mean Lease Obligations by Year
Year N Lease Lease_def
2001 857 256.4 0.196
2002 909 256.1 0.201
2003 915 261.8 0.191
2004 917 271.7 0.186
2005 889 296.4 0.191
2006 859 328.3 0.192
2007 809 379.6 0.194
2008 731 433.1 0.189
2009 707 464.8 0.184
2010 668 486.5 0.195
2011 644 515.7 0.190
2012 640 535.3 0.184
2013 657 529.7 0.182
2014 632 549.5 0.187
2015 600 631.6 0.202
Total 11,434 394.7 0.191
Panel B: Mean Lease Obligation by Industry
Industry N Lease Lease_def
Mining and construction 127 119.7 0.079
Food 201 90.3 0.090
Textiles, printing and publishing 551 196.8 0.140
Chemicals 186 231.4 0.087
Pharmaceuticals 562 40.5 0.105
Extractive industries 170 613.5 0.079
Durable manufacturers 1,708 93.7 0.094
Computers 509 79.9 0.081
Transportation 798 1076.9 0.237
Utilities 61 100.1 0.103
Retail 2,771 903.7 0.361
Services 3,579 168.1 0.153
Other 211 138.5 0.229
Total 11,434 394.7 0.191
The table reports descriptive statistics based on 11,434 firm-year observations between 2001 and 2015
for our lease (Compustat) sample. Lease equals the future lease obligation discounted by a firm-specific
discount rate, as detailed in Appendix A. Lease_def equals Lease deflated by total assets at the end of
the year. Panel A reports means by year. Panel B reports means by industry. We define industry
following Barth et al. (1998).
40
Table 2 Lease Investment Model Panel A: Variable Descriptives
Variable Mean Std Dev P1 Q1 Median Q3 P99
Lease_change 0.013 0.056 -0.099 -0.009 0.001 0.019 0.274
V/P 0.511 0.389 -0.161 0.260 0.440 0.685 1.949
Leverage 0.645 1.351 0.000 0.003 0.233 0.718 6.205
Cash 0.248 0.302 0.001 0.043 0.146 0.353 1.300
Age 2.406 0.978 0.000 1.792 2.485 3.091 4.394
Size 6.299 1.547 3.114 5.204 6.187 7.329 10.202
Stock Returns 1.276 0.890 0.207 0.812 1.103 1.468 5.062
Lease_change, t-1 0.024 0.077 -0.102 -0.007 0.005 0.032 0.349
SalesGr 0.130 0.278 -0.513 0.000 0.090 0.222 1.133
Panel B: Variable Correlations
Lease_
change V/P Leverage Cash Age Size Stock Returns Lease_
change, t-1 SalesGr
Lease_change -0.059 *** -0.042 *** 0.006 -0.079 *** -0.029 *** 0.099 *** 0.294 *** 0.231 *** V/P -0.028 *** 0.077 *** -0.374 *** 0.202 *** 0.217 *** -0.219 *** -0.070 *** -0.108 *** Leverage 0.001 0.230 *** -0.208 *** 0.061 *** 0.244 *** -0.015 -0.043 *** -0.064 *** Cash -0.059 *** -0.467 *** -0.563 *** -0.323 *** -0.351 *** 0.180 *** 0.191 *** 0.130 *** Age -0.011 0.256 *** 0.195 *** -0.297 *** 0.361 *** -0.052 *** -0.168 *** -0.163 *** Size 0.083 *** 0.252 *** 0.471 *** -0.382 *** 0.355 *** -0.147 *** -0.061 *** -0.051 *** Stock Returns 0.148 *** -0.241 *** -0.013 0.057 *** 0.017 * -0.054 *** 0.065 *** 0.179 *** Lease_change, t-1 0.218 *** -0.046 *** -0.018 ** -0.005 -0.104 *** 0.026 *** 0.027 *** 0.366 *** SalesGr 0.278 *** -0.144 *** -0.083 *** 0.061 *** -0.145 *** -0.030 *** 0.199 *** 0.364 ***
41
Table 2 (Continued)
Lease Investment Model
Panel C: Model Estimation
𝐿𝑒𝑎𝑠𝑒_𝑐ℎ𝑎𝑛𝑔𝑒𝑖,𝑡
= 𝛼 + 𝛽1𝑉/𝑃𝑖,𝑡−1 + 𝛽2𝐿𝑒𝑣𝑖,𝑡−1 + 𝛽3𝐶𝑎𝑠ℎ𝑖,𝑡−1 + 𝛽4𝐴𝑔𝑒𝑖,𝑡−1 + 𝛽5𝑆𝑖𝑧𝑒𝑖,𝑡−1
+ 𝛽5𝑅𝑒𝑡𝑢𝑟𝑛𝑖,𝑡−1 + 𝛽6𝐿𝑒𝑎𝑠𝑒_𝑐ℎ𝑎𝑛𝑔𝑒𝑖,𝑡−1 + 𝛽7𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡−1 + 𝑒𝑖𝑡
Variable Coefficient b
V/P -0.0090
(-6.54)
Lev -0.0014 (-3.79)
Cash -0.0116 (-6.03)
Age -0.0020 (-3.65)
Size -0.0008 (-2.36)
Return 0.0036 (6.36)
Lease_changet-1 0.1639 (24.48)
SalesGrt-1 0.0229 (12.14)
Industry Fixed Effects Yes
Year Fixed Effects Yes
Adjusted R2 0.139
The sample covers 12,577 firm years with available data on Compustat for the period 2001–2015.
In Panel A, P1 (P99) is the 1st (99th) percentile of the respective distribution while Q1 (Q3) is the lower (upper)
quartile of the respective distribution. Lease_change equals Lease less lagged Lease, divided by lagged total assets.
V/P equals the ratio of the value of the firm (V) and market value of equity (csho * prcc_f). We estimate V as (1–α
r)BV+ α (1+r)X – α rd where, α equals(ω/(1+r–ω)); r equals 12%; and ω equals 0.62. ω is the abnormal earnings
persistence parameter from the Ohlson (1995) framework, BV is the book value of common equity (CEQ), d is
annual dividends (dvc) and X is operating income after depreciation (oiadp). Lev equals sum of the book value of
short term (DLC) and long term debt (DLTT) deflated by the sum of the book value of total debt and the book value
of equity (CEQ). Cash equals the balance of cash and short term investments (CHE) deflated by total assets
measured at the start of the year. Age equals log of the number of years the firm has been listed on CRSP as of the
start of the year or the number of years on Compustat, if not available. Size equals log of total assets (AT)
measured at the start of the year. Return equals the change in market value of the firm over that prior year
(MVE/MVEt-1). Lease_changet-1 equals Lagged Lease_change. SalesGr equals the change in sales deflated by
beginning total assets. In Panel B, Pearson (Spearman) correlations are reported above (below) the diagonal. T-
statistics are reported in parentheses.
42
Table 3
Future Performance Tests
Panel A: Descriptive Statistics
Variable Mean Std Dev P1 Q1 Median Q3 P99
Lease 394.7 1371.7 1.7 19.8 68.0 231.7 5462.6
Sale1_chg 0.014 0.056 -0.098 -0.009 0.001 0.020 0.275
IB1_chg 0.088 0.234 -0.554 -0.015 0.069 0.183 0.854
IB 0.011 0.123 -0.322 -0.020 0.008 0.033 0.457
SalesGr 0.012 0.158 -0.650 -0.011 0.044 0.090 0.306
ACC 0.102 0.246 -0.545 -0.007 0.078 0.199 0.927
Size -0.074 0.088 -0.366 -0.110 -0.065 -0.027 0.139
Lease_p 6.398 1.543 3.272 5.308 6.296 7.416 10.359
Lease_r 0.013 0.020 -0.019 0.000 0.009 0.022 0.083
ihat_p 0.000 0.050 -0.113 -0.021 -0.006 0.011 0.211
ihat_r 0.080 0.071 -0.028 0.034 0.061 0.108 0.326
43
Table 3, continued
Panel B: Correlations
Sale1_chg IB1_chg Lease_r Lease_p IB SalesGr ACC Size Ihat_p Ihat_r
Sale1_chg 0.173 *** 0.151 *** 0.172 *** 0.154 *** 0.355 *** 0.065 *** -0.030 *** 0.004 0.098 *** IB1_chg 0.316 *** -0.061 *** -0.021 ** -0.306 *** -0.074 *** -0.384 *** -0.052 *** -0.015 -0.049 *** Lease_r 0.135 *** -0.018 * 0.018 * 0.085 *** 0.201 *** -0.019 ** 0.028 *** -0.001 0.186 *** Lease_p 0.195 *** -0.047 *** -0.231 *** 0.137 *** 0.354 *** -0.022 ** -0.048 *** 0.002 0.002 IB 0.265 *** -0.167 *** 0.112 *** 0.271 *** 0.274 *** 0.410 *** 0.307 *** -0.442 *** -0.132 *** SalesGr 0.419 *** -0.012 0.141 *** 0.376 *** 0.417 *** 0.123 *** -0.006 0.090 *** 0.141 *** ACC 0.039 *** -0.235 *** -0.021 ** 0.008 0.271 *** 0.118 *** 0.066 *** -0.096 *** -0.051 *** Size -0.018 * -0.030 *** 0.102 *** -0.014 0.223 *** 0.003 0.054 *** -0.462 *** 0.022 ** Ihat_p 0.040 *** -0.015 0.009 -0.026 *** -0.114 *** 0.159 *** -0.079 *** -0.485 *** 0.127 *** Ihat_r 0.113 *** -0.029 *** 0.173 *** 0.003 -0.008 0.105 *** -0.047 *** 0.108 *** 0.002
The sample covers 11,434 firm years with available data on Compustat for the period 2001–2015. Sale1_chg equals one-year-ahead sales less current sales,
divided by total assets. IB1_chg equals one-year-ahead IB less current IB, divided by total assets. Lease equals future lease obligations discounted with firm-
specific discount rate (as detailed in appendix A); Lease_change equals lease less lagged Lease, divided by lagged total assets. Lease_r equals the residual value
from equation (1) is the estimate of the residual lease change. Lease_p equals the fitted value from equation (1) is the estimate of the expected lease change.
IB equals income before extraordinary items, divided by lagged total assets. SalesGr equals sales less lagged sales, divided by lagged total assets. ACC equals
income before extraordinary items less operating cash flows (OANCF), divided by lagged total assets. Size equals natural log of lagged total assets. ). ihat_r
(ihat_p) is the estimate of the residual (expected) capital expenditures, equaling the residual (expected) value from equation (2). In Panel A, P1 (P99) is the 1st
(99th) percentile of the respective distribution while Q1 (Q3) is the lower (upper) quartile of the respective distribution. In Panel B, Pearson (Spearman)
correlations are reported above (below) the diagonal.
44
Table 3, continued
Panel C: Future Performance Tests – Sales and Earnings Growth
𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+1
= 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝐼𝐵𝑖,𝑡 + 𝛽4𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽5𝐴𝐶𝐶𝑖,𝑡 + 𝑒𝑖𝑡+1
𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+1
= 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝑖ℎ𝑎𝑡_𝑟𝑖,𝑡 + 𝛽4𝑖ℎ𝑎𝑡_𝑝𝑖,𝑡 + 𝛽5𝐼𝐵𝑖,𝑡
+ 𝛽6𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽7𝐴𝐶𝐶𝑖,𝑡 + 𝑒𝑖𝑡+1
Sale1_chg IB1_chg Sale1_chg IB1_chg
Lease_r 0.400 -0.143 0.359 -0.104
(8.50) (-5.37) (7.64) (-4.01)
Lease_p 0.683 -0.143 0.713 0.001
(4.62) (-1.18) (4.69) (0.01)
ihat_r 0.136 -0.115 (5.42) (-6.41)
ihat_p -0.014 -0.277 (-0.29) (-7.37)
IB 0.101 -0.174 0.117 -0.221
(5.17) (-9.16) (5.59) (-11.03) SalesGr 0.287 0.026 0.278 0.043
(16.40) (3.79) (15.57) (6.05) ACC -0.010 -0.425 -0.012 -0.417
(-0.30) (-14.81) (-0.36) (-15.51)
SIZE -0.006 0.002 -0.007 0.000 (-3.40) (2.76) (-3.66) (0.20)
Year Indicators Yes Yes Yes Yes Industry Indicators Yes Yes Yes Yes Adjusted R2 0.198 0.196 0.200 0.215 The sample covers 11,434 firm years with available data on Compustat for the period 2001–2015. t-statistics are
clustered by firm and reported in parentheses. We define industry following Barth et al. (1998). Year (industry)
Indicators is a vector of indicator variables to capture annual (industry) fixed effects.
45
Table 3, continued
Panel D: Future Performance Tests – Sales and Earnings Growth within Subsamples of Unexpected Lease Changes
𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+1 = 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝐼𝐵𝑖,𝑡 + 𝛽4𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽5𝐴𝐶𝐶𝑖,𝑡 + 𝛽6𝑆𝐼𝑍𝐸𝑖,𝑡 + 𝑒𝑖𝑡+1
𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+1 = 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝑖ℎ𝑎𝑡_𝑟𝑖,𝑡 + 𝛽4𝑖ℎ𝑎𝑡_𝑝𝑖,𝑡 + 𝛽5𝐼𝐵𝑖,𝑡 + 𝛽6𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽7𝐴𝐶𝐶𝑖,𝑡 + 𝑒𝑖𝑡+1
The second, fourth, sixth and eighth (third, fifth, seventh and ninth) columns contain results for firm-years in which Lease_r was greater than (less than) zero.
We define industry following Barth et al. (1998). Year and Industry Indicators are vectors of indicator variables to capture annual and industry fixed effects.
t-statistics are clustered by firm and reported in parentheses.
Sales Change Earnings Change
Lease_r>0 Lease_r<0 Lease_r>0 Lease_r<0 Lease_r>0 Lease_r<0 Lease_r>0 Lease_r<0
Lease_r 0.249 0.606 0.224 0.565 -0.165 -0.160 -0.142 -0.109
(3.67) (4.31) (3.29) (4.04) (-4.39) (-2.15) (-3.85) (-1.49)
Lease_p 0.722 0.916 0.772 0.942 -0.032 -0.285 0.080 -0.123
(2.99) (3.93) (3.15) (3.93) (-0.28) (-1.61) (0.67) (-0.68)
ihat_r 0.093 0.177 -0.098 -0.135
(2.76) (4.71) (-4.25) (-4.85)
ihat_p -0.034 -0.018 -0.234 -0.293
(-0.46) (-0.28) (-5.40) (-5.96)
IB 0.130 0.077 0.135 0.099 -0.120 -0.203 -0.152 -0.260
(4.37) (3.03) (4.46) (3.58) (-4.69) (-8.64) (-5.88) (-10.17)
SalesGr 0.278 0.289 0.269 0.281 0.011 0.038 0.028 0.055
(11.91) (12.08) (11.20) (11.56) (1.29) (4.12) (3.03) (5.71)
ACC -0.037 0.006 -0.033 -0.003 -0.283 -0.511 -0.293 -0.488
(-0.75) (0.14) (-0.68) (-0.08) (-7.63) (-13.44) (-8.04) (-14.08)
SIZE -0.007 -0.006 -0.008 -0.007 0.001 0.003 -0.001 0.001
(-2.73) (-2.80) (-2.91) (-3.00) (0.81) (2.64) (-0.61) (0.54)
Year Indicators Yes Yes Yes Yes Yes Yes Yes Yes
Industry Indicators Yes Yes Yes Yes Yes Yes Yes Yes
Observations 4,575 6,859 4,575 6,859 4,575 6,859 4,575 6,859
Adjusted R2 0.221
0.173 0.223 0.176 0.118 0.242 0.137 0.260
46
Table 4 Association between Leases and Two- and Three-year-ahead Performance
2 𝑜𝑟 3 − 𝑦𝑒𝑎𝑟 𝐹𝑢𝑡𝑢𝑟𝑒 𝑃𝑒𝑟𝑓𝑜𝑟𝑚𝑎𝑛𝑐𝑒𝑖,𝑡+2,3
= 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝑖ℎ𝑎𝑡_𝑟𝑖,𝑡 + 𝛽4𝑖ℎ𝑎𝑡_𝑝𝑖,𝑡 + 𝛽5𝐼𝐵𝑖,𝑡 + 𝛽6𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽7𝐴𝐶𝐶𝑖,𝑡 + 𝛽8𝑆𝑖𝑧𝑒𝑖,𝑡 + 𝑒𝑖𝑡+2,3
lease_r>0 lease_r>0 lease_r>0 lease_r>0
Sale2_chg IB2_chg Sale3_chg IB3_chg Sale2_chg IB2_chg Sale3_chg IB3_chg
Lease_r 0.465 -0.070 0.398 -0.078 0.224 -0.128 0.289 -0.112
(4.74) (-2.06) (2.52) (-1.94) (1.54) (-2.48) (1.17) (-1.80)
Lease_p 2.069 0.075 3.155 0.294 2.344 -0.056 3.764 0.120
(6.36) (0.51) (6.05) (1.62) (4.69) (-0.31) (4.58) (0.51)
ihat_r 0.133 -0.150 0.251 -0.105 -0.043 -0.088 0.005 -0.077
(2.65) (-6.77) (2.97) (-3.69) (-0.71) (-3.05) (0.05) (-1.96)
ihat_p 0.043 -0.295 0.045 -0.373 0.130 -0.208 0.419 -0.264
(0.37) (-6.00) (0.23) (-5.70) (0.79) (-3.33) (1.72) (-3.26)
IB 0.252 -0.271 0.400 -0.273 0.262 -0.161 0.474 -0.180
(5.46) (-9.73) (5.20) (-8.26) (4.14) (-4.16) (4.80) (-3.87)
SalesGr 0.435 0.032 0.608 0.014 0.419 0.018 0.611 0.018
(11.66) (3.11) (10.53) (1.21) (8.24) (1.43) (7.72) (1.14)
ACC -0.024 -0.478 -0.121 -0.505 -0.113 -0.335 -0.328 -0.390
(-0.35) (-14.28) (-1.09) (-12.84) (-1.14) (-6.87) (-2.08) (-6.66)
Size -0.019 0.002 -0.037 0.002 -0.018 -0.001 -0.025 0.001
(-4.33) (1.75) (-4.92) (1.17) (-3.07) (-0.48) (-2.42) (0.29)
Industry Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Firm Fixed Effects Yes Yes Yes Yes Yes Yes Yes Yes Adjusted R2 0.170 0.214 0.161 0.188 0.193 0.125 0.201 0.117
The sample covers 9,998 (8,751) firm years with available data for the two-year-ahead (three-year-ahead) tests. Sale2_chg (Sale3_chg) equals the two(three)-
year-ahead sales less current sales, divided by total assets. IB2_chg (IB3_chg) equals two(three)-year-ahead IB less current IB, divided by total assets.
Remaining variables are defined in Table 3. t-statistics, clustered by firm, are shown in parentheses below the coefficients.
47
Table 5
Tests of the Association between Leases and Contemporaneous Stock Returns
Panel A: Descriptive statistics
Variable Mean Std Dev P1 Q1 Median Q3 P99
Hret_1 0.136 0.654 -0.846 -0.200 0.069 0.342 2.319
Earn_p 0.010 0.144 -0.516 0.002 0.038 0.062 0.176
B_P 0.505 0.359 0.053 0.255 0.424 0.657 1.765
Accruals -0.033 0.059 -0.193 -0.063 -0.033 -0.004 0.138
Invest 0.049 0.107 -0.124 -0.006 0.023 0.074 0.490
Noach 0.060 0.139 -0.229 -0.017 0.034 0.110 0.563
Panel B: Correlations
Hret_1 Earn_p B_P Accruals Invest Noach
Hret_1
-0.077 *** 0.129 *** -0.020 *** -0.044 *** -0.060 ***
Earn_p 0.105 ***
-0.098 *** 0.148 *** 0.097 *** 0.174 ***
B_P 0.120 *** 0.193 ***
-0.024 *** -0.053 *** -0.060 ***
ACCR -0.027 *** 0.162 *** -0.022 ***
0.226 *** 0.415 ***
Invest -0.063 *** 0.184 *** -0.075 *** 0.268 ***
0.705 ***
Noach -0.077 *** 0.177 *** -0.073 *** 0.456 *** 0.745 ***
Panel C: Estimation of the Penman and Yehuda (2016) Model for Expected Returns
𝑅𝑡 = 𝛼 + 𝛽1𝐸𝑎𝑟𝑛𝑖𝑛𝑔𝑠𝑡−1
𝑃𝑡−1+ 𝛽2
𝐵𝑡−1
𝑃𝑡−1+ 𝛽3𝐴𝐶𝐶𝑅𝑡−1 + 𝛽4𝐼𝑁𝑉𝐸𝑆𝑇𝑡−1 + 𝛽5∆𝑁𝑂𝐴𝑡−1 + 휀𝑡
Intercept Earnings/Pt-1 B t-1/P t-1 ACCR t-1 INVEST t-1 ΔNOA t-1 Adj. R2
Coefficient: 0.100 0.071 0.087 0.027 -0.122 -0.123 0.044
t-statistic: 2.44 1.42 2.62 0.26 -0.52 -0.91
The sample covers 29,305 firm years with available data for the period 1998–2015. Hret_1 equals stock return for
year t, calculated as buy‐and‐hold compounded monthly returns from CRSP over the period from three months
after the beginning of the fiscal year t to three months after the end. This is the period during which accounting
data for fiscal year is reported. Earn_p equals earnings for fiscal‐year t before extraordinary items (Compustat item
IB) and special items (item SPI), minus preferred dividends (item DVP), with a tax allocation to special items at the
prevailing federal statutory corporate income tax rate for the year. Earningst-1/Pt‐1 is the earnings yield for fiscal
year t. B_P equals book value of common equity divided by market value of equity. Accruals equals accruals
divided by average total assets. Accruals is measured as the sum of change in accounts receivable (item RECT),
change in inventory (item INVT), and change in other current assets (item ACO), minus the sum of change in
accounts payable (item AP) and change in other current liabilities (item LCO), minus depreciation and amortization
expense (item DP). Invest equals investment calculated as (change in gross property, plant, and equipment (item
PPENT) + change in inventory (item INVT) + change in off-balance-sheet leases)/ lagged assets. Noach equals the
change in net operating assets (after capitalizing off-balance-sheet leases) divided by average total assets. In Panel
A, P1 (P99) is the 1st (99th) percentile of the respective distribution while Q1 (Q3) is the lower (upper) quartile of
the respective distribution. In Panel B, Pearson (Spearman) correlations are reported above (below) the diagonal.
48
Table 5, continued
Tests of the Association between Leases and Contemporaneous Stock Returns
Panel D: Tests of Returns 𝐶𝑆𝐴𝑅𝑖,𝑡 = 𝛼 + 𝛽1𝐿𝑒𝑎𝑠𝑒_𝑟𝑖,𝑡 + 𝛽2𝐿𝑒𝑎𝑠𝑒_𝑝𝑖,𝑡 + 𝛽3𝑖ℎ𝑎𝑡_𝑟𝑖,𝑡 + 𝛽4𝑖ℎ𝑎𝑡_𝑝𝑖,𝑡 + 𝛽5𝐸𝑥𝑝𝑟𝑒𝑡_𝑐ℎ𝑔𝑖,𝑡 + 𝛽6𝐼𝐵𝑖,𝑡
+ 𝛽7𝑆𝑎𝑙𝑒𝑠𝐺𝑟𝑖,𝑡 + 𝛽8𝐴𝐶𝐶𝑖,𝑡 + 𝛽9𝑟𝐵𝑇𝑀𝑖,𝑡−1 + 𝛽10𝑟𝑀𝑉𝐸𝑖,𝑡−1 + 𝑒𝑖𝑡
Model: (1) (2) (3) (4)
Lease_r>0
Lease_r -0.078 -0.420 -0.375 -0.535 (-0.62) (-3.26) (-2.92) (-2.31)
Lease_p -2.525 -0.671 -0.650 -0.082
(-6.02) (-1.50) (-1.46) (-0.09) ihat_r -0.291 -0.389
(-3.91) (-3.74) ihat_p -0.274 -0.302
(-1.98) (-1.18) Expret_chg -1.580 -1.643 -2.047
(-10.80) (-10.93) (-6.91) IB 1.065 0.627 0.564 0.760
(8.35) (5.66) (5.17) (3.28) SalesGr 0.349 0.330 0.355 0.310
(8.41) (8.19) (8.69) (5.03) ACC -0.522 -0.593 -0.637 -0.676
(-5.61) (-6.53) (-6.94) (-4.70) rBTM 0.147 0.171 0.153 0.133
(8.28) (9.63) (7.85) (4.16) rMVE -0.044 -0.024 -0.031 -0.045
(-2.54) (-1.41) (-1.72) (-1.43) Year Indicators Yes Yes Yes Yes Industry Indicators Yes Yes Yes Yes Observations 7,091 7,091 7,091 2,911 Adjusted R2 0.094 0.141 0.145 0.159
CSAR equals annual size-adjusted abnormal returns, beginning on the first day of the fourth month of the current
fiscal year and extending to the last day of the third month after the fiscal year end. Lease_r (Lease_p) is the
estimate of the residual (expected) lease change, equaling the residual (expected) value from equation (1). ihat_r
(ihat_p) is the estimate of the residual (expected) capital expenditures, equaling the residual (expected) value from
equation (2). Expret_chg equals change in expected return from Penman and Yehuda (2016) model. IB equals
income before extraordinary items, divided by lagged total assets. SalesGr equals sales less lagged sales, divided by
lagged total assets. ACC equals Income Before Extraordinary Items less Operating Cash Flows (OANCF) , divided by
lagged total assets. rBTM equals the yearly quintile rank for beginning of the year Book value of shareholders’
equity divided by market value of equity, standardized to range between 0 and 1. rMVE equals the yearly quintile
rank of beginning of the year market value of equity, standardized to range between 0 and 1. T-statistics are
reported in parentheses.
49
Table 6
Tests of the Association between Leases and Executive Compensation
Panel A: Descriptive Statistics
Variable Mean Std Dev P1 Q1 Median Q3 P99
TotalCompi,t+1 4,715.710 4,921.670 257.270 1,580.430 3,172.660 6,056.310 25,145.910
Sale i,t 5,327.380 13,270.950 73.702 594.769 1,476.070 4,151.200 77,946.000
ROA i,t 0.055 0.086 -0.262 0.025 0.057 0.098 0.273
MTB i,t 3.278 3.024 0.784 1.623 2.436 3.709 18.988
RET i,t 15.896 48.529 -70.960 -14.066 9.901 36.916 189.785
ROA_vol i,t 0.047 0.053 0.003 0.015 0.027 0.055 0.261
RET_vol i,t 51.071 42.422 9.149 26.012 39.516 60.288 250.474
DRES i,t 0.452 0.498 0.000 0.000 0.000 1.000 1.000
Dihat i,t 0.371 0.483 0.000 0.000 0.000 1.000 1.000
Panel B: Compensation Tests
𝐶𝑂𝑀𝑃𝑖,𝑡+1 = 𝛼 + 𝛽1𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝑅𝑂𝐴𝑖,𝑡 + 𝛽3𝑀𝑇𝐵𝑖,𝑡 + 𝛽4𝑅𝐸𝑇𝑖,𝑡 + 𝛽5𝑅𝑂𝐴_𝑣𝑜𝑙𝑖,𝑡 + 𝛽6𝑅𝐸𝑇_𝑣𝑜𝑙𝑖,𝑡−1
+ 𝑒𝑖,𝑡+1
𝐶𝑂𝑀𝑃𝑖,𝑡+1 = 𝛼 + 𝛽1𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝐷𝑅𝐸𝑆𝑖,𝑡 + 𝛽3𝐷𝑅𝐸𝑆_𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽4𝑅𝑂𝐴𝑖,𝑡 + 𝛽5𝑀𝑇𝐵𝑖,𝑡 + 𝛽6𝑅𝐸𝑇𝑖,𝑡
+ 𝛽7𝑅𝑂𝐴_𝑣𝑜𝑙𝑖,𝑡 + 𝛽8𝑅𝐸𝑇_𝑣𝑜𝑙𝑖,𝑡−1 + 𝑒𝑖𝑡+1
𝐶𝑂𝑀𝑃𝑖,𝑡+1 = 𝛼 + 𝛽1𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝐷𝑅𝐸𝑆𝑖,𝑡 + 𝛽3𝐷𝑅𝐸𝑆_𝑆𝑎𝑙𝑒𝑠𝑖,𝑡 + 𝛽2𝐷𝑖ℎ𝑎𝑡𝑖,𝑡 + 𝛽3𝐷𝑖ℎ𝑎𝑡_𝑆𝑎𝑙𝑒𝑠𝑖,𝑡
+ 𝛽4𝑅𝑂𝐴𝑖,𝑡 + 𝛽5𝑀𝑇𝐵𝑖,𝑡 + 𝛽6𝑅𝐸𝑇𝑖,𝑡 + 𝛽7𝑅𝑂𝐴_𝑣𝑜𝑙𝑖,𝑡 + 𝛽8𝑅𝐸𝑇_𝑣𝑜𝑙𝑖,𝑡−1 + 𝑒𝑖𝑡+1
Model: (1) (2) (3) (4) Lease_r>0
Sales 0.140 0.120 0.162 0.188 (23.92) (17.53) (14.26) (17.51) DRES 10.055 -77.763 (0.06) (-0.47) DRES_Sales 0.066 0.077
(5.43) (6.23) Dihat 619.152 (3.67) Dihat_Sales -0.062 (-4.80) ROA 1814.707 1765.377 1830.413 2219.756
(1.88) (1.84) (1.91) (1.46) MTB 138.771 133.098 129.995 148.657
(5.21) (5.02) (4.92) (3.71) RET 7.953 7.694 7.578 6.437 (4.69) (4.55) (4.50) (2.68) ROA_vol -4575.534 -4241.356 -3932.646 -4103.494
(-2.83) (-2.63) (-2.45) (-1.58) RET_vol -5.788 -6.021 -5.796 -8.189
(-2.99) (-3.12) (-3.01) (-2.67) Year Indicators Yes Yes Yes Yes Industry Indicators Yes Yes Yes Yes Observations 3,456 3,456 3,456 1,561 Adjusted R2 0.227 0.234 0.240 0.247
50
TotalComp equals one-year-ahead total compensation as reported in SEC filings, which equals the sum of SALARY,
BONUS, STOCK_AWARDS, OPTION AWARDS, NONEQ_INCENT, PENSION_CHG, and OTHCOMP. Sale is revenues for
the year prior to the year in which compensation is awarded. DRES equals 1 if lease_r is greater than zero, and 0
otherwise. DRES_Sale is an interaction term equaling Sale multiplied by DRES. Dihat equals 1 if ihat_r is greater
than zero, and 0 otherwise. Dihat_Sale is an interaction term equaling Sale multiplied by Dihat. MTB equals the
market to book ratio averaged over the five years ended the year prior to the year in which CEO compensation was
paid. ROA equals the ratio of income before extraordinary items to average total assets for the prior year. Stock
return is the percentage stock market return for the prior year. ROA_vol is the standard deviation of ROA for the
prior five years. RET_vol is the standard deviation of annual percentage stock market return for the prior five
years. T-statistics are reported in parentheses.
51
Table 7
Unexplained Investments in Leases and Capital Expenditures, and Free Cash Flows
Panel A: Descriptive Statistics by Lease Residual Quintile
Lease Residual Quintile 1 (n=2,280)
Differences: Quintile 5 – Quintile 1
Variable Mean Median Mean Median
Lease_r -0.050 -0.040 0.119 *** 0.084 ***
Ihat_r -0.017 -0.027 0.049 *** 0.029 ***
FCF -0.022 0.002 0.035 *** 0.022 ***
ROA -0.015 0.036 0.045 *** 0.017 ***
Sale_chg 0.082 0.076 0.126 *** 0.084 ***
Lease Residual Quintile 2 (n=2,291)
Variable Mean Median
Lease_r -0.017 -0.017
Ihat_r -0.009 -0.019
FCF -0.007 0.013
ROA -0.006 0.040
Sale_chg 0.082 0.065
Lease Residual Quintile 3 (n=2,287)
Variable Mean Median
Lease_r -0.006 -0.006
Ihat_r -0.009 -0.018
FCF 0.000 0.018
ROA 0.002 0.040
Sale_chg 0.064 0.053
Lease Residual Quintile 4 (n=2,291)
Variable Mean Median
Lease_r 0.007 0.007
Ihat_r -0.001 -0.012
FCF 0.009 0.022
ROA 0.015 0.043
Sale_chg 0.083 0.063
Lease Residual Quintile 5 (n=2,285)
Variable Mean Median
Lease_r 0.068 0.044
Ihat_r 0.032 0.002
FCF 0.012 0.024
ROA 0.029 0.053
Sale_chg 0.208 0.160
52
Table 7, continued
Panel B: Descriptive Statistics by Free Cash Flow Quintile
Free Cash Flow Quintile 1 (n=2,280)
Differences: Quintile 5 – Quintile 1
Variable Mean Median Mean Median
Lease_r -0.008 -0.011 0.012 *** 0.006 ***
Ihat_r -0.022 -0.040 0.036 *** 0.034 ***
FCF -0.194 -0.144 0.340 *** 0.268 ***
Free Cash Flow Quintile 2 (n=2,291)
Variable Mean Median
Lease_r 0.001 -0.005
Ihat_r -0.006 -0.024
FCF -0.032 -0.030
Free Cash Flow Quintile 3 (n=2,287)
Variable Mean Median
Lease_r 0.003 -0.004
Ihat_r 0.003 -0.012
FCF 0.016 0.015
Free Cash Flow Quintile 4 (n=2,291)
Variable Mean Median
Lease_r 0.001 -0.004
Ihat_r 0.007 -0.007
FCF 0.057 0.056
Free Cash Flow Quintile 5 (n=2,285)
Variable Mean Median
Lease_r 0.004 -0.004
Ihat_r 0.014 -0.006
FCF 0.146 0.124
Panel C: Relation between over-investment (lease_r and ihat_r) and free cash flow (FCF)
𝐿𝑒𝑎𝑠𝑒_𝑟 = 𝛼 + 𝛿1𝐹𝐶𝐹 < 0 + 𝛿2𝐹𝐶𝐹 > 0 + 휀
𝐼ℎ𝑎𝑡_𝑟 = 𝛼 + 𝛿1𝐹𝐶𝐹 < 0 + 𝛿2𝐹𝐶𝐹 > 0 + 휀
Model α δ1 δ2 F-statistic for test δ1 = δ2
Lease_r 0.0010 0.0285 0.0164 1.44
(1.56) (5.94) (2.18)
Ihat_r -0.0077 0.0083 0.1705 77.65***
(6.51) (0.95) (12.42)
53
The sample covers 11,434 firm years with available data on Compustat for the period 2001–2015. Lease_r equals
the residual value from equation (1), the estimate of the residual lease change. ihat_r is the estimate of the
residual capital expenditures, equaling the residual value from equation (2). FCF equals free cash flows, defined as
operating cash flows less depreciation and amortization plus research and development, deflated by average total
assets less expected investment from equation (2). ROA equals income before extraordinary items, divided by
average total assets. Sale_chg equals current sales less lagged sales, deflated by beginning of year assets.