Leases and Over-investment†

Tim Eaton Miami University

eatont@miamioh.edu

Craig Nichols Syracuse University dcnichol@syr.edu

James Wahlen

James R. Hodge Chair of Excellence Indiana University

jwahlen@indiana.edu

Matthew Wieland Miami University

wielanmm@miamioh.edu

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.