Amortized Cost for Operating Lease Assets - American Accounting ...

Accounting Horizons Vol. 27, No. 1 2013 pp. 51–74

American Accounting Association DOI: 10.2308/acch-50278

Amortized Cost for Operating Lease Assets Ross Jennings and Ana Marques SYNOPSIS: A proposed accounting standard issued jointly by the Financial Accounting Standards Board (FASB) and the International Accounting Standards Board (IASB) would require firms to recognize many more lease assets than are currently required and to amortize those assets on a straight-line basis. A number of respondents to the exposure draft argue that the ‘‘front-loading’’ of lease expense resulting from straight-line amortization would not reflect the economics of the lease assets. This study compares straight-line amortization with the most-often cited alternative, present value amortization. First, we illustrate by example that under stylized conditions, present value amortization provides information that more faithfully represents the future cash flows of lease assets than straight-line amortization. Second, for a large subset of firms that are more likely to conform to the stylized conditions in our example, we find that investors value those firms as though the lease assets are capitalized and amortized on a present value basis. Finally, we find that financial ratio comparability is substantially increased when operating leases are constructively capitalized and amortized using straight-line amortization, and further increased when using present value amortization. Taken together, these results provide no evidence for favoring straight-line amortization over present value amortization as the default method for amortizing capitalized operating leases. Keywords: leasing; straight-line amortization; present value amortization; value-relevance. Data Availability: Data used in this paper are publicly available.

INTRODUCTION

T

his study compares two alternatives for computing amortized cost for operating lease assets: straight-line amortization and present value amortization. This issue is important and timely because it is currently being considered by the Financial Accounting Standards

Ross Jennings is a Professor at The University of Texas at Austin, and Ana Marques is an Assistant Professor at Nova School of Business and Economics. We acknowledge the excellent research assistance of Arash Aloosh. Professor Jennings gratefully acknowledges the financial support of the PwC Centennial Fellowship and the McCombs School of Business. Professor Marques gratefully acknowledges the financial support of the Foundation for Science and Technology (PTDC/ECO/72021/2006) and INOVA.

Submitted: November 2011 Accepted: May 2012 Published Online: August 2012 Corresponding author: Ross G. Jennings Email: [email protected]

51

52

Jennings and Marques

Board (FASB) and the International Accounting Standards Board (IASB). The Boards issued a joint exposure draft for a new leasing standard in August 2010, and plan to issue a second exposure draft in 2012. The proposed standard would require lessees to record a ‘‘right-of-use asset’’ and associated liability when entering into a noncancellable lease arrangement, bringing large amounts of currently unrecognized assets and liabilities onto corporate balance sheets. Initial measurement for both the asset and the liability would be at the present value of expected future lease payments. Subsequent to initial recognition, the liability would be reduced using the effective interest method and the asset would be amortized using a method that reflects the pattern of consumption of economic benefits, with straight-line amortization as the default method if an alternative pattern cannot be reliably determined. Relative to current standards, use of straight-line amortization for lease assets would accelerate expense recognition, lowering net income during the early part of the lease’s life and on a cumulative basis. In their public comments on the exposure draft, a number of respondents argue that the ‘‘front-loading’’ of lease expense that would result from straight-line amortization will reduce comparability across firms, because two companies with identical leased assets who are at different years in the lease term, but are otherwise identical, will report different net income. To avoid this outcome, to promote comparability, and to provide information that more faithfully represents the future cash flows of these lease arrangements, some respondents have recommended setting amortization of the lease asset equal to the reduction in principle of the lease liability, so that total lease expense would be equal to the lease payment, mitigating or eliminating the effect on net income of the new standard. This can be accomplished by utilizing a form of present value amortization in place of straight-line amortization. While this argument appears to be self-serving—to avoid the reduction in income the proposed standard would produce for firms whose operating lease activity is growing—it is nevertheless supported by a long line of theoretical research in accounting and economics, beginning with Hotelling (1925). In this study, we examine the Boards’ choice of straight-line amortization as the default method for amortizing lease assets. We conduct this analysis in two stages. In the first stage, we use stylized examples to compare the two alternative amortization methods. In the second stage, we use a sample of 34,707 firm-year observations drawn from 19 industries with substantial operating leases for the years 1998–2008 to compare the two alternative amortization methods in terms of investor valuations and their effect on financial statement comparability. In the first stage, we conclude that whether alternative amortization methods provide relevant and representationally faithful information to financial statement users depends on both the amount and pattern of the expected future cash flows. In particular, under the assumption that lease assets are zero net present value investments with certain future cash inflows, present value amortization dominates straight-line amortization in providing representationally faithful information about the future cash flows of the lease asset (Anton 1956; Feltham and Ohlson 1996). In contrast, when lease assets are positive net present value investments, neither present value amortization nor straight-line amortization clearly dominates the other in providing information that is representationally faithful about future cash flows. In the second stage, we first regress stock prices for sample firms on balance sheet components, including the incremental asset that would be recorded if lease assets were valued on a present value basis rather than a straight-line basis. For the sample as a whole, our results provide evidence that investors view lease assets, on average, as positive net present value investments. We then repeat the analysis for two subsamples of firms that vary on the basis of relative profit margin within their industry. For firms in the higher profitability group, we find that investors view lease assets, on average, as positive net present value investments. For firms in the lower profitability group, the Accounting Horizons March 2013

Amortized Cost for Operating Lease Assets

53

results are consistent with zero net present value lease assets that are amortized using present value amortization. Finally, we assess the effect of capitalizing operating leases and choice of amortization method on financial statement comparability, using three common financial ratios. We find that capitalizing operating leases substantially improves comparability regardless of amortization method, and that the improvement is slightly greater using present value amortization. Taken together, the results in this study provide no support for straight-line amortization over present value amortization as the default method for amortizing capitalized operating leases. Rather, our analyses indicate that operating leases are positive net present value investments for many firms for which neither method is clearly favored over the other. In addition, for firms that are most likely to have zero net present value leases, we find some evidence that investors value those firms consistently with present value amortization. Finally, we find evidence that present value amortization enhances comparability slightly more than straight-line amortization. The next section summarizes prior research and comments on the proposed standard for lease accounting. The third section describes a simple example that illustrates the difference between straight-line amortization and present value amortization. The fourth section presents our main empirical analysis, and concluding comments are provided in the fifth section. BACKGROUND AND THE RESEARCH QUESTION Prior Research Research articles and textbooks have advocated ‘‘constructively capitalizing’’ operating leases for many years, and generally agree that lease assets and liabilities should be recorded at the present value of expected future lease payments, and the liabilities should be reduced over time using the effective interest method (e.g., Imhoff et al. 1997; White et al. 2003; Stickney et al. 2007; Wild et al. 2007; Easton et al. 2009; Revsine et al. 2009). There is less agreement on how to value the lease asset over time. Some estimate the lease asset as equal to the lease liability (White et al. 2003; Stickney et al. 2007; Wild et al. 2007; Easton et al. 2009), which is implicitly consistent with amortization on a present value basis. Others explicitly amortize the lease asset on a straight-line basis (Imhoff et al. 1997; Revsine et al. 2009). A number of studies provide evidence on the materiality of capitalizing operating leases for financial statement analysis in the U.S. (Imhoff et al. 1991), in the U.K. (Beattie et al. 1998), in New Zealand (Bennett and Bradbury 2003), and in Canada (Durocher 2008). An additional line of research focuses on the effect on measures of leverage and risk of omitting operating lease assets and liabilities from the balance sheet (Imhoff et al. 1993; Ely 1995; Boatsman and Dong 2011; Dhaliwal et al. 2011). The Proposed Change to Lease Accounting In 2010, the Financial Accounting Standards Board and the International Accounting Standards Board jointly issued an exposure draft for a revised accounting standard for lease accounting (FASB 2010). The Exposure Draft (ED) states that the objective of the proposed standard is ‘‘to report relevant and representationally faithful information to users of financial statements about the amounts, timing, and uncertainty of the cash flows arising from leases’’ (FASB 2010, para. 4). To accomplish this, the ED requires lessees to record a liability and asset at commencement of the lease that are both equal to the present value of expected future lease payments. Subsequently, the liability is reduced as payments are made, using the effective interest method. Requirements for amortizing the lease asset are provided in FASB (2010, para. 20): Accounting Horizons March 2013


54

A lessee shall amortize the right-of-use asset on a systematic basis from the date of commencement of the lease to the end of the lease term or over the useful life of the underlying asset if shorter. The lessee shall select the amortization method . . . in accordance with Topic 350. Topic 350 provides the following guidance in 350-30-35-6: A recognized intangible asset shall be amortized over its useful life to the reporting entity unless that life is determined to be indefinite . . . The method of amortization shall reflect the pattern in which the economic benefits of the intangible asset are consumed or otherwise used up. If that pattern cannot be reliably determined, a straight-line amortization method shall be used. Thus, the ED requires straight-line amortization unless ‘‘the pattern in which the economic benefits of the intangible asset are consumed’’ can be ‘‘reliably determined’’ by the reporting entity (FASB 2010).1 Relative to current practice, straight-line amortization for leased assets would accelerate expense recognition, resulting in lower net income in the asset’s early years and cumulatively over the asset’s life. Reaction to the Proposed Changes Predictably, many respondents to the ED reacted negatively to this outcome.2 For example, in a typical letter, McDonald’s made the following comment: The proposal significantly changes the pattern of income/expense recognition from both the lessee and lessor perspective with the front loading of income/expense. We believe that the proposed approach will provide less relevant information to users of our financial statements and does not reflect the economics of the transactions. In their letter, Starbucks argued that: Recognizing higher occupancy cost in the early part of the lease compared to the latter part is out of step with the underlying economics and would misrepresent net income. It would not enhance a user’s ability to make comparisons among companies, and would significantly complicate the ability to understand trends or forecast future results. Finally, Toys‘R’Us argues that financial statement users will have to undo the effects of this mismatching to make the financial statements more useful: We feel these drastic changes in expense recognition will jeopardize the reliability and comparability of our financial information, forcing financial statement users and investors to continue adjusting our results to compensate for the front loading effect of the proposed guidance. Thus, respondents argue that straight-line amortization for lease assets will provide lessrelevant information that does not faithfully represent the economics of leasing arrangements, and would reduce comparability across firms, requiring investors to continue adjusting financial 1

2

These requirements are consistent with the Boards’ preliminary view on this issue presented in FASB (2009a, para. 5.19). The Boards also considered valuing the lease assets at fair value, but decided to value them at cost to be consistent with the measurement of other nonfinancial assets and to minimize complexity (FASB 2009a, paras. BC72–74). However, at the same time, the Boards concluded that ‘‘cost usually provides a reasonable approximation to the fair value of the right-of-use asset at its inception’’ (FASB 2009a, para. BC72). For a summary, see FASB (2011, paras. 67–68). Similar comments were made in response to the earlier Discussion Paper; see FASB (2009b, paras. 44–46).

Accounting Horizons March 2013


55

statements for weaknesses in lease accounting.3 The primary alternative amortization method proposed by respondents to the ED is present value amortization, also referred to as ‘‘economic,’’ ‘‘annuity,’’ and ‘‘mortgage-based’’ amortization. Under this method, amortization of the lease asset is equal to the decline in value from one period to the next in the present value of the remaining expected future lease payments, and total lease expense is equal to the lease payment. Applying this method would eliminate the net income effect of the new standard. Amortized Cost Alternatives Present value amortization was considered by the Boards as part of the so-called ‘‘linked approach,’’ under which ‘‘a lease that does not have the same economic effect as a loan and purchase should be accounted for in a way that links the amortization of the asset and liability arising from the lease’’ (FASB 2010, para. BC8(b)). The Boards rejected this alternative for three reasons described in paragraph BC10 of the ED. First, the Boards decided that this alternative was not consistent with the treatment for other financial liabilities, even though the liability treatment is not different under this approach from the approach adopted by the Boards. Second, the Boards rejected this alternative because they viewed that the values of the lease asset and liability should not be necessarily linked subsequent to inception of the lease. However, subjecting the lease asset to subsequent review for impairments, as the proposed standard requires (FASB 2010, para. 24), results in asset and liability values that are not necessarily linked. Third, the Boards rejected this alternative because ‘‘the treatment of amounts recognized in the income statement is inconsistent with the boards’ view that the lessee has acquired a right-of-use asset and is paying for that right over the term of the lease’’ (FASB 2010, BC10(c)). However, the only necessary distinction between straight-line amortization and present value amortization is the amounts that remain on the balance sheet and flow through the income statement as expense, not their consistency with the right-of-use model. During re-deliberation of the exposure draft at a joint Board meeting on February 28–29, 2012, the FASB and IASB considered a third alternative amortization method, the underlying asset approach (FASB 2012). This method is a blend of the straight-line and present value methods, and produces a result that lies between the two. The pattern of amortization produced by the underlying asset approach (UAA) varies with the portion of the underlying asset consumed during the lease period, and is identical to straight-line (present value) amortization when all (none) of the remaining life of the underlying asset is consumed during the lease period. In this study, we compare the two alternative methods for applying amortized cost to lease assets, straight-line amortization and present value amortization. We do not explicitly examine UAA amortization because doing so would require us to estimate for each firm-year the average portion of the underlying asset expected to be consumed during the lease period. However, because UAA amortization is itself a weighted-average of straight-line amortization and present value amortization, our results also allow for some inferences about UAA amortization. Both straight-line amortization and present value amortization have strengths and weaknesses. The principle strengths of the straight-line method are that it is simple to use and is commonly used for other finite-lived intangible assets. Thus, it is not surprising that this is the method preferred by standard setters and many others. However, there is little economic rationale for straight-line amortization. In contrast, the present value method, although less familiar to most practicing accountants, is also relatively simple to use and, as applied to depreciating physical assets, has a 3

Note that although ‘‘comparability’’ is not stated as an objective of the proposed standard in FASB (2010) paragraph 4, it is mentioned in the basis for conclusions in 14 separate paragraphs (BC6, BC10, BC72, BC74, BC77, BC83, BC84, BC94, BC125, BC137, BC182, BC188, BC189, and BC204).



56

long history dating back to a famous paper by Hotelling (1925). Anton (1956) uses examples to compare the pattern of present value depreciation to other common depreciation methods for various patterns of future cash flows, and concludes in favor of present value depreciation. Present value depreciation for physical assets was also the subject of an active discussion in the accounting literature in the 1960s.4 More recently, Feltham and Ohlson (1996) discuss present value depreciation as having ‘‘desirable’’ properties under conditions of zero net present value investments with certain future cash flows.5 The intuition underlying present value depreciation for physical assets, as applied to lease assets, is that otherwise-equivalent lease assets of different ages should earn about the same rate of return. If shorter leases of the same asset earn a higher rate of return than longer leases, the price of the shorter leases would be bid up until the rates of return are equalized. Conditional on estimates of the amount and timing of expected future cash flows, present value amortization computes the decline in value of the lease asset in such a way that the rate of return on the asset is equal over the life of the lease, an outcome that will be achieved by straight-line amortization only if the expected future cash flows follow a certain pattern. In contrast, present value amortization insures that the lease’s rate of return is equal for each year of its life for any pattern of expected future cash flows, as long as the lease asset is zero net present value and there is certainty about the amount and timing of the future cash flows (Feltham and Ohlson 1996). In the next section, we present a simple example of capitalizing an operating lease to demonstrate (1) the differences between straight-line amortization and present value amortization, and (2) the importance of whether the lease is a zero or positive net present value investment. In the following section, we use a large sample of firms that rely on leasing to compare the two amortization methods in terms of investor valuation and financial statement comparability. EXAMPLES OF ALTERNATIVE AMORTIZATION METHODS The Basic Example Assuming Zero Net Present Value We illustrate the difference between straight-line amortization and present value amortization using the following stylized example (presented in Table 1), which is consistent with the conditions identified by Feltham and Ohlson (1996), zero net present value and certain future cash flows. Assume a five-year lease with an option to renew for a second five years that is initially recorded as a ten-year lease with annual payments of $1,000.6 The proposed accounting standard requires that at inception of the lease, an asset and liability be recorded that are both equal to the present value of the expected lease payments discounted at the lessee’s incremental borrowing rate (FASB 2010, para. 12(a)). In this example, we assume a discount rate of 7 percent so that the initial present value is $7,024, and under the proposed accounting in the exposure draft, the company will record an asset and a liability for this amount. In subsequent years, the liability will be reduced by the amount of the annual payment less the accrued interest expense for the current period, and the asset will decline according to its amortization schedule (subject to potential impairments, which is a separate issue and beyond the scope of this study). 4

5

6

In this discussion, authors generally favored using present value depreciation for physical assets, but differed in some details of application. See, for example, Coughlan (1959), Dixon (1960), Reynolds (1961), Bodenhorn (1961), Bierman (1961), and Staubus (1962). The ‘‘desirable’’ properties are book value equal to market value, accounting earnings equal to economic earnings, and book rate of return equal to cost of capital. We demonstrate these features in our example below. In this and the following examples, for computational convenience, we assume that all cash inflows and outflows occur at the end of each period. This assumption is not necessary for any of the conclusions we reach; rather, we only need to assume that cash inflows and outflows follow the same pattern within each period (e.g., at beginning or end of period or evenly throughout the period) for both amortization methods.



57

TABLE 1 Example of Alternative Amortization Methods for a Lease Asset Panel A: Zero NPV Asset Straight-Line Amortization Year of Lease 1

3

5

7

9

Cash Inflow Straight-Line Amortization Income Before Interest

1,000 702 298

1,000 702 298

1,000 702 298

1,000 702 298

1,000 702 298

Beginning Book Value SL Amort Rate of Return Beg MV Asset/Beg BV Asset

7,024 4.2% 1.00

5,619 5.3% 1.06

4,214 7.1% 1.13

2,809 10.6% 1.21

1,404 21.2% 1.29

Panel B: Zero NPV Asset Present Value Amortization Year of Lease 1

3

5

7

9

Cash Inflow Present Value Amortization Income Before Interest

1,000 508 492

1,000 582 418

1,000 666 334

1,000 763 237

1,000 873 127

Beginning Book Value PV Amort Rate of Return Beg MV Asset/Beg BV Asset

7,024 7.0% 1.00

5,971 7.0% 1.00

4,767 7.0% 1.00

3,387 7.0% 1.00

1,808 7.0% 1.00

Panel C: Positive NPV Asset Straight-Line Amortization Year of Lease 1

3

5

7

9

Cash Inflow Straight-Line Amortization Income Before Interest

1,500 702 798

1,500 702 798

1,500 702 798

1,500 702 798

1,500 702 798

Beginning Book Value SL Amort Rate of Return Beg MV Asset/Beg BV Asset (SL)

7,024 11.4% 1.50

5,619 14.2% 1.59

4,214 18.9% 1.70

2,809 28.4% 1.81

1,404 56.8% 1.93

Panel D: Positive NPV Asset Present Value Amortization Year of Lease 1

3

5

7

9

Cash Inflow Present Value Amortization Income Before Interest

1,500 508 992

1,500 582 918

1,500 666 834

1,500 763 737

1,500 873 627

Beginning Book Value PV Amort Rate of Return Beg MV Asset/Beg BV Asset (SL)

7,024 14.1% 1.50

5,971 15.4% 1.50

4,767 17.5% 1.50


3,387 1,808 21.8% 34.7% 1.50 1.50 (continued on next page)


58

TABLE 1 (continued)

This table presents four examples of alternative combinations of (1) the pattern of cash inflows over the life of a ten-year lease, and (2) the amortization of the lease asset. In the first two cases, the present value of the future cash inflows generated by the lease is $7,024, so that the lease is a zero net present value investment. In the last two cases, the present value of the future cash inflows generated by the lease is $10,535, so that the lease is a positive net present value investment. In all four cases, net income is cash inflow minus amortization, and rate of return is net income divided by beginning book value of the lease asset. In Panel A, the cash inflows are constant at $1,000 each year for ten years, and straight-line amortization each year is 10 percent ($702) of the original lease value. In Panel B, the cash inflows are constant at $1,000 each year for ten years, and present value amortization each year is equal to the decline in present value of the remaining future cash flows from beginning to end of period. In Panel C, the cash inflows are constant at $1,500 each year for ten years, and straight-line amortization each year is 10 percent ($702) of the present value of the lease payments. In Panel D, the cash inflows are constant at $1,500 each year for ten years, and present value amortization each year is equal to the decline in present value of the remaining future lease payments from beginning to end of period.

Table 1 presents two alternative amortization schedules for this lease asset. The first is straightline amortization and is presented in Panel A. In this example, we assume the lease asset is a zero net present value investment, with net cash inflows that are equal in amount and timing to the lease payments and that are discounted at the same incremental borrowing rate of 7 percent. In this example, straight-line amortization is $702.40 per year for each of the ten years. In this case, total expense from the lease is greater than the lease payment in the early years of the lease and less in the later years. For example, total lease expense in the first (last) year of the lease’s life is the sum of interest expense of $491.65 ($65.42) and amortization expense of $702.40 ($702.40), which is equal to $1,194.05 ($767.82) and greater (less) than the lease payment of $1,000. Moreover, the rate of return on the lease asset grows over time, from 4.2 percent in the first year to 21.2 percent in the ninth year. As noted as far back as Hotelling (1925), this is unlikely to reflect the underlying economic value of the leased asset, where market forces should insure that the asset generates a stable rate of return over its life. The final line of Panel A reports the ratio of the market value of the lease asset to its book value.7 This ratio is 1.00 at inception of the lease, but slowly rises over the life of the lease as the book value of the lease asset under straight-line amortization falls more rapidly than its market value. By the ninth year, this ratio is 1.29, indicating that the asset’s book value is substantially below its market value. The second alternative is present value amortization, under which the asset value each year is the present value of the remaining cash payments, and the value of the lease asset declines at the same rate as the lease liability. This is represented in Panel B of Table 1, where we continue to assume that the amount and timing of the cash inflows are equal to the lease payments. In this case, amortization expense is lower in the early years of the asset’s life and higher in later years, and total lease expense equals the lease payment each year.8 More importantly, the rate of return is the same 7

8

Throughout the examples illustrated in Table 1, we assume that the market has the same expectations of the future cash inflows from the lease as the lessee, and uses the lessee’s incremental borrowing rate to discount the expected future cash inflows. We also assume that those shared expectations do not change over time and are fully realized each period, consistent with the certainty condition in Feltham and Ohlson (1996). For example, under present value amortization, total lease expense in the first (last) year of the lease’s life is the sum of interest expense of $492 ($65) and amortization expense of $508 ($935), which is equal to $1,000 ($1,000), the amount of the lease payment. This example is similar to Case I in Anton (1956).



59

in every year, equal to the discount rate of 7 percent. The last line in Panel B reports that the market value and book value of the lease asset are equal each year over the life of the lease. In Appendix A, we apply the underlying asset approach (UAA) to amortization of lease assets to this example. This extension demonstrates that UAA amortization is a weighted-average of straight-line amortization and present value amortization, and will be equal to straight-line (present value) amortization if, during the lease, all (none) of the remaining life of the underlying asset is consumed. Thus, under this method, a lease that is for the entire life of the asset and is, essentially, a purchase of the asset, will be amortized on a straight-line basis, while a lease that is for land, in which none of the asset’s remaining life is consumed, will be amortized on a present value basis.9 Taken together, these results indicate that for this stylized example, present value amortization provides information that faithfully represents the future cash flows of the lease asset, and is comparable across firms that have similar leased assets with remaining leases of different lengths. In contrast, this is generally not the case for straight-line amortization (as shown in Panel A of Table 1) or for the underlying asset approach (as shown in the Appendix A). Feltham and Ohlson (1996, 209–210) describe these characteristics (book value equal to market value and stable rate of return) as potentially ‘‘desirable’’ outcomes of present value depreciation. A stable rate of return and book value equal to market value is possible with straight-line amortization, but only if the future net cash inflows are expected to follow a particular pattern of decline over the life of the lease.10 In contrast, any pattern of net cash inflows that has a present value of $7,024 will produce a stable rate of return and book value equal to market value under present value amortization. An additional feature of present value amortization (Panel B) is that this method does not require gains and losses to be recorded when expectations about the term or payments for the lease are subsequently changed. This is especially important because under the proposed standard, both the term and the amount of future payments may be uncertain estimates at the inception of the lease. In the example above, if, at the end of the first five years, the lessee decides not to renew the lease, the remaining lease liability and asset would need to be reversed. Under straight-line amortization (Panel A), the value of the lease liability is greater than the value of the lease asset at the end of the fifth year (and throughout the life of the lease), so that failure to renew would result in the lessee recording a one-time gain. In contrast, under present value amortization (Panel B), the liability and the asset have the same value, so there would be no gain or loss for reversing the lease asset and liability. The Extended Example Assuming Positive Net Present Value All of the above discussion assumes that leased assets are zero net present value investments; i.e., that the present value of the cash inflows and the present value of the lease payments are equal when both are discounted using the lessee’s incremental borrowing rate. This is a reasonable assumption for highly competitive markets. However, it is possible that in some markets, lessees are able to negotiate lease terms that permit them to earn abnormal profits on their lease assets, so that the present value of the cash inflows is greater than the present value of the lease payments. We examine this possibility next, in Panels C and D of Table 1, where the original example is adjusted to result in a positive net present value investment by increasing the annual net cash inflows from $1,000 to $1,500, while keeping the lease payments at $1,000 per year. This results in a market 9

10

Further examples of this approach are available in IASB (2012a, 2012b). In IASB (2012a), the airplane (building) example is based on the unstated assumption that the eight- (ten-) year lease begins when there is 18 (45) years remaining in the asset’s life. In IASB (2012b), the equipment example is based on the unstated assumption that the five-year lease begins when there is ten years remaining in the asset’s life. Examples of this can be seen in Anton (1956, Case III) and Feltham and Ohlson (1996, Appendix A.2).



60

value for the leased asset that is greater than the book value, both at acquisition and throughout the lease term, regardless of which amortization method is used. The pattern of results in Panel C, for straight-line amortization, is similar to the pattern of results reported in Panel A for a zero net present value investment. However, when the lease is positive net present value, the increase in the rate of return over the life of the lease is more dramatic and the ratio of the market value of the lease asset to its book value is larger, 1.93 in the ninth year in Panel C, compared to 1.29 in the ninth year in Panel A. Thus, the distortion between the accounting for the lease asset and the underlying economics is greater when the lease is a positive net present value investment rather than zero net present value. The pattern of results in Panel D, for present value amortization, contrasts with the pattern of results in Panel B. In the positive net present value case, the rate of return for the lease asset is rising over time, much as with straight-line amortization, and the market value of the lease asset is a constant multiple of its book value. Thus, in this case, the market value of the asset exceeds the book value regardless of whether straight-line or present value amortization is used, although the distortion is less in both magnitude and variability for present value amortization. However, this suggests that relative to zero net present value leases, for positive net present value leases, both alternatives provide information that is less comparable across firms and less faithfully represents the future cash flows of the lease asset.11 Taken together, the results for the simple examples presented in Table 1 indicate that under ‘‘ideal’’ conditions, when the lease asset is a zero net present value investment, present value amortization produces information that is more comparable and that more faithfully represents the future cash flows of the lease asset. On the other hand, when the lease asset is a positive net present value investment, both alternatives provide less useful information. Although the ideal conditions assumed in the examples are very restrictive, the examples demonstrate that, in general, the closer that lease arrangements are to these ideal conditions, the more likely present value amortization will provide more useful information about lease assets than straight-line amortization. Moreover, there are no conditions under which straight-line amortization provides more useful information than present value amortization.12 In the next section, we provide empirical evidence on how these two amortization methods relate to investor valuation, and how they affect financial statement comparability across firms within industries. EMPIRICAL ANALYSIS The Sample We begin our sample selection process with all firms for which there are data in the Compustat database for the period from 1998 to 2008. We require all observations to have data for our main variables of analysis (defined below): EPS, Price, Inc_Adj, A, L, LA_Adj, and Profit_Mg. This results in a sample of 41,753 observations.13 11

12

13

An alternative implementation of present value amortization that results in a stable rate of return and book value equal to market involves using the internal rate of return, rather than the incremental borrowing rate, as the discount rate and recording a gain at inception for positive net present value leases. However, this alternative was rejected by the Boards in their decision to value the lease asset at cost rather than fair value (see FASB 2010, para. BC72). This conclusion can be extended to the underlying asset approach (UAA) to amortization. As we demonstrate in the Appendix A, UAA amortization is a weighted-average of straight-line and present value amortization, so that as with straight-line amortization, there are no conditions under which UAA amortization provides more useful information than present value amortization. This initial sample excludes 305 observations with unusually high average expected lives for their lease assets of more than 45 years based on our algorithm (see below).



61

Next, we control for influential observations by removing 3,356 observations in the 1 percent tails for the variables mentioned above. Finally, we divide the observations into industry groups on the basis of four-digit GICS (Global Industry Classification Standard) codes, and we identify industries for which operating leases are a substantial part of the operations for the average firm in the industry. We retain only industries with at least 500 firm-year observations and for which the average ratio of future lease payments to total assets is at least 0.05. This results in a final sample of 34,707 observations in 19 different industries. Table 2 presents for each industry the number and percent of firm-year observations, the average ratio of total future expected lease payments to total assets, the average expected life of a leased asset, and the lease assets to lease liabilities ratio (ALR, discussed below). Three industries each contribute more than 10 percent of the observations in our sample: software and services, technology hardware and equipment, and capital goods. The retailing industry has the largest ratio of expected future lease payments to total assets (58.3 percent), followed by consumer services (38.7 percent), and transportation (24.9 percent). The average expected life for lease assets is the number of years for which we have expected future cash flows following the algorithm described below. This length varies from 12.42 years for the real estate industry to 6.22 years for the software and services industry. Variable Construction In our empirical analysis, we focus on the balance sheet rather than the income statement. Although respondents to the exposure draft comment more on the income statement effects of straight-line amortization than the balance sheet effects, empirically, the balance sheet effects are more material and easier to estimate. This is because straight-line amortization records more expense than present value amortization in the early part of the lease and less expense in the later part of the lease, and the two expense amounts are approximately equal at the midpoint of the lease. In contrast, the dollar difference in asset values is near its peak at the midpoint of the lease, and any estimation error will be only in amount, not in direction.14 We estimate the value of off-balance sheet lease assets (LA) and lease liabilities (LL) for sample firms based on disclosures of the minimum lease payments the firm is currently required to make for each of the next five fiscal years (Pmttþ1 . . . Pmttþ5), as well as a lump sum for all years beyond the fifth year (PmtT). Following Imhoff et al. (1997), we assume that the expected payment in the fifth year continues in the sixth year and beyond, until all of the ‘‘thereafter’’ payments are accounted for. We use a constant interest rate across the sample (7 percent) to discount the expected future lease payments, and this present value is our estimate of the off-balance sheet lease liability (LL) that will be added to the balance sheet. We construct two alternative specifications for the off-balance sheet lease asset. The first is based on present value amortization and is equal to the lease liability (LA_PV ¼ LL). The second is based on straight-line amortization (LA_SL). For this estimate, we recognize that the value of the lease asset will decline more quickly after inception of the lease than the value of the lease liability, and we compute an observation-specific estimate of the ratio between the lease asset and lease liability (ALR) as follows: 1 ð1 þ rÞN Rem N r : ALR ¼ Rem 1 ð1 þ rÞ r

14

Using the examples in Panels A and B of Table 1 to illustrate these points, in year 5 lease expense under straightline amortization is $1,036, 3.6 percent higher than present value amortization, while in year 6 lease expense under straight-line amortization is $989, 1.1 percent lower than present value amortization. In contrast, in year 5 the lease asset is $588 higher on a present value basis and is still $578 higher in year 6.



62

TABLE 2 Observations for Each Industry and Average Ratio of Future Expected Lease Payments to Total Assets SIC

Industry

Obs.

%

P/Assets

Life

ALR

1510 2010 2020 2030 2510 2520 2530 2540 2550 3020 3030 3510 3520 4020 4040 4510 4520 4530 5010

Materials Capital Goods Commercial and Prof. Services Transportation Automobiles and Components Consumer Durables and Apparel Consumer Services Media Retailing Food, Beverage, and Tobacco Household and Personal Products Health Care Eq. and Services Pharm., Biotech., and Life Sciences Diversified Financials Real Estate Software and Services Technology Hardware and Eq. Semiconductors and Semicond. Eq. Telecommunication Services

1,971 3,702 1,659 693 579 1,998 1,558 1,276 1,935 1,063 579 3,159 2,683 1,127 705 3,862 3,760 1,487 911

5.68 10.67 4.78 2.00 1.67 5.76 4.49 3.68 5.58 3.06 1.67 9.10 7.73 3.25 2.03 11.13 10.83 4.28 2.62

0.055 0.065 0.155 0.249 0.052 0.093 0.387 0.136 0.583 0.064 0.124 0.128 0.136 0.111 0.057 0.152 0.092 0.057 0.114

7.53 6.97 6.88 8.15 6.84 6.54 9.97 8.03 8.58 7.50 6.25 6.44 6.70 7.23 12.42 6.22 6.33 6.92 7.80

0.891 0.898 0.898 0.883 0.899 0.903 0.864 0.884 0.877 0.892 0.907 0.905 0.902 0.894 0.847 0.907 0.906 0.899 0.889

Total

34,707

This table presents, for each industry and for the full sample, the number of firm-year observations (Obs.) and the percent of total sample observations (%). Variable Definitions: P/Assets ¼ average ratio of future lease payments to total assets; Life ¼ average expected life of a lease. It is computed as the number of years for which each firm-year discloses future lease payments, assuming that the payment to be made in the fifth out-year continues until all subsequent payments are accounted for; and ALR ¼ average ratio of the leased asset to the leased liability. It is computed as the formula below indicates. In this formula, N is the number of years for which we have expected future cash flows, which is equal to 5 þ PmtT/Pmttþ5 rounded up to the next integer, and Rem ¼ N/2: 1 ð1 þ rÞN Rem N r : ALR ¼ 1 ð1 þ rÞRem r

The numerator of this ratio (the relative asset value) is equal to the present value of a one-dollar annuity for the original life of the lease (N years) times the ratio of the remaining years (Rem) to the original life of the lease (N). The denominator of this ratio (the relative liability value) is equal to the present value of a one-dollar annuity for the remaining years of the life of the lease (Rem). For this calculation, we assume that the original life of the lease (N) is the number of years for which we have expected future cash flows (5 þ PmtT/Pmttþ5 ) rounded up to the next integer. We assume that Accounting Horizons March 2013


63

the remaining years of the life of the lease (Rem) is one-half of its original life (Rem ¼ N/2).15 Thus, the value of the lease asset that is added to the balance sheet under straight-line amortization is given by LA_SL ¼ LL ALR. Average values for ALR are reported for each industry in the final column of Table 2. This ratio varies from 0.847 for real estate to 0.907 for software and services and household and personal products. Allowing the value of the lease asset to differ from the value of the lease liability implies that both net income and deferred tax assets will be affected. We assume that the expense for operating leases for tax purposes is the amount of the lease payment, while the expense for reporting purposes is the sum of the interest expense portion of the payment and the current period amortization of the lease asset. This accelerates operating lease expense for reporting purposes relative to operating lease expense for tax purposes and results in a deferred tax asset, which we denote as LDTA ¼ (LL LA_SL)(T), where T is the statutory tax rate, which we assume to be 35 percent, the current maximum statutory tax rate for corporations.16 Regressions of Security Price on Balance Sheet Components We use the following regression each year to estimate the valuation multiple for the incremental lease asset that would be recorded under present value amortization:17 Price ¼ c0 þ c1 ðA þ LA SL þ LDTAÞ þ c2 ðL þ LLÞ þ c3 LA Adj þ e;

ð1Þ

where Price is the price per share of common equity on the last day of the current fiscal year, A (L) is equal to recorded assets (liabilities), and LL, LA_SL, and LDTA are as defined above. Our primary variable of interest, LA_Adj, is equal to (LA_PV (LA_SL þ LDTA)). This is the additional asset that would be recorded if present value amortization were used, rather than straight-line amortization. All variables are on a per-share basis. The average across the sample for LA_Adj is 11.24 cents per share.18 In Regression (1), we expect a positive coefficient estimate for assets and a negative coefficient estimate for liabilities. In both cases, the theoretical value for the coefficient is 1.19 Our main focus is on the coefficient estimate for LA_Adj. If investors view lease assets as zero net present value investments and value them on a straight-line basis, the coefficient estimate on this variable should 15

16

17

18

19

This assumption explicitly assumes that the average lease asset is at the midpoint of its expected useful life. This differs from textbook presentations in White et al. (2003), Wild et al. (2007), and Easton et al. (2009), where the lease asset is amortized over the number of projected future payments, which implicitly assumes the average lease asset is at the beginning of its expected useful life. This procedure for adjusting the balance sheet for operating leases is generally consistent with the procedure described by Imhoff et al. (1997). Revsine et al. (2009) follow a similar procedure, but ignore the tax effect. Alternative versions of valuation regressions that include as independent variables only balance sheet components (such as Regression (1)) have been used commonly in prior research to address a wide range of accounting issues. Examples from earlier research include studies of pension accounting (Barth 1991), fair value accounting for banks (Barth 1994; Nelson 1996; Eccher et al. 1996), inventory accounting (Jennings et al. 1996), accounting for post-employment benefits (Choi et al. 1997; Davis-Friday et al. 1999), accounting for deferred taxes (Ayers 1998), and oil and gas accounting (Boone 2002). This analysis explicitly assumes that investors (1) are aware of the information available in the footnotes about estimated lease payments, and (2) use that information to adjust the financial statements as if operating leases were recognized as assets and liabilities. De Franco et al. (2011) provide corroborating evidence of investor attention to footnote information concerning operating leases. The pair-wise correlations between LA_Adj and (A þ LA_SL þ LADT) and (L þ LL) are 0.306 and 0.300, respectively, suggesting the potential for collinearity in the annual estimations of Regression (1). However, statistical tests are based on cross-year standard errors of the annual coefficient estimates, which are not affected by collinearity in the annual regressions. This will be the case if amortized cost (book value) equals market value, and is consistent with zero net present value assets.



64

TABLE 3 Regressions of Security Price on Balance Sheet Components Including the Adjustment to Net Assets for Constructively Capitalizing Operating Leases Using Straight-Line Amortization Rather than Present Value Amortization Regression 1: Including Adjustment for Leases

Avg. Coeff. t-statistic

Intercept

Assets

Liabilities

Asset Adj.

R2

8.00 (14.01)

0.96 (17.92)

0.87 (15.24)

2.16 (3.45)

0.35

This table presents average coefficient estimates for the following regression for the years 1999 to 2008: Price ¼ c0 þ c1 ðA þ LA PV þ LDTAÞ þ c2 ðL þ LLÞ þ c3 LA Adj þ e ð1Þ where Price is the price per share of common equity on the last day of the current fiscal year, A (L) is equal to recorded assets (liabilities), LL is lease liability, and LA_PV is lease asset on a present value basis (equal to LL). LA_Adj is equal to (LL (LA_SL þ LDTA)), where LA_SL is equal to the estimated lease asset on a straight-line basis and LDTA is equal to the deferred tax asset. All variables are on a per-share basis. t-statistics are based on the standard deviation of the distribution of annual coefficient estimates. For the column headings in the table: Assets ¼ A þ LA_PV þ LDTA, Liabilities ¼ L þ LL, and Asset Adj. ¼ LA_Adj.

be 0. In contrast, if investors view the lease assets as zero net present value investments and value them on a present value amortization basis, then the coefficient estimate on this variable should be positive and about equal to 1. Finally, if investors view lease assets as positive net present value investments, then the coefficient estimate on this variable should be greater than 1 regardless of the amortization method implicitly used, consistent with the examples in Panels C and D of Table 1.20 The average results for estimation of Regression (1) each year are reported in Table 3, where tstatistics are based on the cross-year coefficient distribution. The average coefficient estimates for total assets and total liabilities are 0.96 (t ¼ 17.92) and 0.87 (t ¼ 15.24), respectively, and the average R2 is 0.35. These results are consistent with those presented in prior research for similar regressions. Our main focus is the average coefficient estimate for LA_Adj, the net asset difference between accounting for lease assets using present value amortization and accounting for lease assets using straight-line amortization. This average coefficient estimate is 2.16 (t ¼ 3.45), which is significantly greater than both 0 (p-value ¼ 0.003, one-tailed) and 1 (t ¼ 1.86, p-value ¼ 0.046, one-tailed).21 This suggests that investors view lease assets for our sample, on average, as positive net present value investments. To examine this possibility more carefully, we divide the sample into two groups based on whether their profit margin (ratio of net income to net revenue) is above or below the median profit margin for that industry. Observations in the higher (lower) profitability group are earning a higher (lower) return on their overall operations and, by extension, are more likely to have lease assets that are positive (zero) net present value investments. Table 4 reports the median profit margin for each industry, and the median for each profitability subgroup within each industry. We observe substantial within-industry variation in profit margins, consistent with the idea that profit margin may capture meaningful differences in the extent to which 20

21

While we are unable to explicitly test amortization based on the underlying asset approach, results for that approach would be similar to straight-line (present value) amortization to the extent that the portion of the underlying asset consumed during the lease period is high (low). All p-values are two-tailed unless specifically indicated otherwise. One-tailed tests are reported when the alternative hypothesis is directional.



65

TABLE 4 Median Profit Margin for Each Industry and for High and Low Profit Margin Subgroups within Each Industry

SIC

Industry

Median for Industry

1510 2010 2020 2030 2510 2520 2530 2540 2550 3020 3030 3510 3520 4020 4040 4510 4520 4530 5010

Materials Capital Goods Commercial and Prof. Services Transportation Automobiles and Components Consumer Durables and Apparel Consumer Services Media Retailing Food, Beverage, and Tobacco Household and Personal Products Health Care Eq. and Services Pharm., Biotech., and Life Sciences Diversified Financials Real Estate Software and Services Technology Hardware and Eq. Semiconductors and Semicond. Eq. Telecommunication Services Column Median

0.033 0.033 0.022 0.035 0.012 0.030 0.039 0.012 0.025 0.033 0.027 0.017 0.740 0.076 0.102 0.008 0.004 0.029 0.017 0.027

Median for High Profit Margin Group

Median for Low Profit Margin Group

0.073 0.065 0.062 0.078 0.040 0.064 0.082 0.087 0.052 0.064 0.072 0.071 0.036 0.178 0.225 0.083 0.064 0.117 0.106 0.072

0.022 0.009 0.030 0.008 0.049 0.019 0.022 0.148 0.000 0.004 0.088 0.189 3.605 0.041 0.013 0.365 0.171 0.153 0.177 0.041

This table presents, for each industry, the industry median profit margin and the median profit margin for the high and low profit margin subgroups within each industry. The high (low) profit margin observations are above (below) the industry median each year. Profit margin is the ratio of net income to sales revenue for that firm-year observation.

lease investments are positive net present value investments for firms in these industries that make extensive use of operating leases. The median industry profit margin for the high profit margin group is 0.072, while the median for the low profit margin group is 0.041, and these two medians are statistically different (p-value , 0.000). This suggests that lease assets for the high (low) profit margin group may be more likely to be positive (zero) net present value investments, on average.22 Next, we re-estimate a version of Regression (1) each year that permits separate coefficients for each variable, including the intercept, for each profitability group, as follows: Price ¼ Ri¼1;2 ½c0þi Di þ c1þi Di ðA þ LA PV þ LDTAÞ þ c2þi Di ðL þ LLÞ þ c3þi Di ðLA AdjÞ þ e: ð2Þ Di is an indicator variable that takes the value of 1 for observations in the ith (first or second) profitability group, and 0 otherwise. All other variables are as defined above. 22

The low median profit margin for the low profitability group for the pharmaceutical industry (3.605) and, to a lesser extent, for the software and services industry (0.365) is due to early-stage companies with very little revenue.



66

TABLE 5 Regressions of Security Price on Balance Sheet Components Including the Adjustment to Net Assets for Constructively Capitalizing Operating Leases Using Straight-Line Amortization Rather than Present Value Amortization Regression 2: Including Profitability Interactions Intercept

Assets

Liabilities

Asset Adj

Low PM Group Coefficient t-statistic

4.90 (8.19)

0.70 (17.97)

0.60 (12.42)

1.17 (1.67)

High PM Group Coefficient t-statistic

10.31 (17.04)

1.20 (28.74)

1.13 (23.04)

4.67 (4.08)

R2

0.67

This table presents average coefficient estimates for the following regression for the years 1999 to 2008: Price ¼ Ri ¼ 1; 4½c0þi Di þ c1þi Di ðA þ LA PV þ LDTAÞ þ c2þi Di ðL þ LLÞ þ c3þi Di ðLA AdjÞ þ e;

ð2Þ

where Price is the price per share of common equity on the last day of the current fiscal year, A (L) is equal to recorded assets (liabilities), LL is lease liability, and LA_PV is lease asset on a present value basis (equal to LL). LA_Adj is equal to (LL (LA_SL þ LDTA)), where LA_SL is equal to the estimated lease asset on a straight-line basis and LDTA is equal to the deferred tax asset. High (low) profit margin groups are observations that are above (below) the industry median profit margin that year, where profit margin is the ratio of net income to sales revenue. All variables are on a per-share basis. tstatistics are based on the standard deviation of the distribution of annual coefficient estimates. For the column headings in the table: Assets ¼ A þ LA_PV þ LDTA, Liabilities ¼ L þ LL, and Asset Adj. ¼ LA_Adj.

The results are reported in Table 5. For the lower profitability group, the average coefficient estimate for LA_Adj is 1.17, which is marginally statistically greater than 0 (t ¼ 1.67, p-value ¼ 0.063, one-tailed), but not significantly greater than 1 (t ¼ 0.24, p-value ¼ 0.408, one-tailed). This provides weak evidence, consistent with zero net present value lease assets for this group, that investors implicitly value using present value amortization.23 The results are much different for the higher profitability group. For this group, the average coefficient estimate is 4.67, which is statistically greater than both 0 (t ¼ 4.08, p-value ¼ 0.001, onetailed) and 1 (t ¼ 3.21, p-value ¼ 0.005, one-tailed). For this group, the results indicate that investors view lease assets as positive net present value investments.24 The results in this section indicate that there is substantial variation across observations based on profit margin in how investors implicitly value lease assets. For firms in the higher profitability group, investors view lease assets as positive net present value investments, for which neither straight-line amortization nor present value amortization will faithfully represent the underlying economics. For firms in the lower profitability group, there is some evidence suggesting that (1) lease assets are zero net present value investments, and (2) investors implicitly value lease assets 23

24

This evidence is weak in two ways. First, in the statistical sense, the standard error is large relative to the coefficient estimate, providing weak tests of whether the coefficient estimate is different from either 0 or 1. Also, in terms of interpretation, although the results are consistent with investors viewing the lease assets for these firms as 0 net present value projects that are valued on a present value basis, we cannot rule out the alternative that investors view these projects as positive net present value projects that are valued on a straight-line basis. Note that the effect of allowing the regression coefficients to vary between the high and low profitability groups is to raise the average R2 from 0.35 for Regression (1) to 0.67 for Regression (2).



67

for these firms using present value amortization. Overall, this analysis provides no evidence in support of requiring straight-line amortization as the default amortization method for operating leases. Financial Statement Comparability In this section, we provide empirical evidence on the effect on comparability of capitalizing operating leases and using either straight-line amortization or present value amortization. We operationalize comparability using the within-industry standard deviations for three common financial ratios that measure return on assets, asset turnover, and financial leverage. An accounting method that lowers the within-industry standard deviation for an important financial ratio can be said to enhance comparability, because that method results in similar companies (based on common industry membership) appearing more similar (based on a lower within-industry standard deviation). Thus, if capitalizing and amortizing operating leases results in financial statements that are more comparable, we expect the within-industry standard deviation for these three ratios will be lower after constructively capitalizing operating leases on the balance sheet relative to using asreported numbers. Similarly, if one method of amortizing capitalized operating leases results in financial statements that are more comparable than another method, we expect the within-industry standard deviation for these three ratios will be lower when using asset and liability numbers based on the first method rather than the second method. For this analysis, return on assets is the ratio of net income to beginning-of-period total assets, asset turnover is the ratio of net sales to beginning-of-period total assets, and financial leverage is the ratio of beginning-of-period total liabilities to beginning-of-period total assets. To compute the average standard deviation for each industry in the sample, we first compute each ratio for each firm-year in the sample, then compute the standard deviation for each ratio within each industry and year, and then average these standard deviations across years. The results are reported in the first three columns of Table 6. Overall, it is apparent that the within-industry variation for all three ratios declines substantially from the as-reported column to the straight-line column, and then declines further from the straight-line column to the present value column. For example, the average within-industry standard deviation for asset turnover on an asreported basis is 1.842, which falls by 28.2 percent to 1.323 when total assets is computed after capitalizing operating leases and amortizing them on a straight-line basis. We then observe a further reduction by 0.6 percent to 1.315 when total assets is computed after capitalizing operating leases and amortizing them on a present value basis. In the fourth and fifth columns of Table 6, we report the average and median across the 19 industries for the differences in the within-industry standard deviation between as-reported and straight-line amortization and present value amortization for capitalized operating leases, respectively. We also report the frequency of positive and negative differences across the 19 industries, and the associated p-values based on a binomial distribution and the null hypothesis of equal probability for positive and negative differences. For all three ratios, the within-industry standard deviations are reduced for at least 18 of the 19 industries for both straight-line amortization and present value amortization, and this is significant at less than the 0.001 level in all six cases.25 Most importantly, the final column in Table 6 compares the within-industry standard deviation for ratios based on amortizing operating leases using straight-line amortization with ratios based on amortizing operating leases using present value amortization. We find that even though the differences are generally small, the within-industry standard deviations are reduced for at least 18 of 25

The only case in which the within-industry standard deviation did not decline for straight-line amortization, relative to as-reported, is for financial leverage in the real estate industry, which increased from 0.215 to 0.352.



68

TABLE 6 Within-Industry Standard Deviation for Three Common Financial Ratios As-Reported, Using Straight-Line Amortization, and Using Present Value Amortization

Ratio Return on Assets Average Median Pos. Differences Neg. Differences p-value Asset Turnover Average Median Pos. Differences Neg. Differences p-value Financial Leverage Average Median Pos. Differences Neg. Differences p-value

Diff. As-Reported versus SL

Diff. As-Reported versus PV

Diff. SL versus PV

0.563 0.571

0.582 0.126 19 0 ,0.001

0.587 0.131 19 0 ,0.001

0.004 0.003 19 0 ,0.001

1.323 1.120

1.315 1.115

0.519 0.174 19 0 ,0.001

0.527 0.202 19 0 ,0.001

0.008 0.006 18 1 ,0.001

0.467 0.410

0.447 0.407

0.149 0.102 18 1 ,0.001

0.168 0.094 19 0 ,0.001

0.020 0.004 18 1 ,0.001

As-Reported Avg. Std. Dev.

SL Avg. Std. Dev.

PV Avg. Std. Dev.

1.149 0.687

0.567 0.574

1.842 1.269

0.616 0.518

This table presents average and median within-industry standard deviations and differences in within-industry standard deviations for three financial ratios. The average within-industry standard deviation for each industry ratio is computed by first computing the standard deviation for each ratio for each industry each year, and then averaging across years. Return on Assets is net income divided by beginning-of-period total assets. Asset Turnover is net sales divided by beginning-ofperiod total assets. Financial Leverage is beginning-of-period total liabilities divided by beginning-of-period total assets. The first column reports the average and median within-industry standard deviation for each ratio across the 19 industries in the sample on an as-reported basis. The second (third) column reports the average and median within-industry standard deviation for each ratio across the 19 industries in the sample assuming that operating leases are capitalized and amortized on a straight-line (present value) basis. The fourth (fifth) column reports the average and median across the 19 industries for the differences in the withinindustry standard deviation between as-reported and straight-line (present value) amortization for capitalized operating leases, the frequency of positive and negative differences for the 19 industries, and the associated p-values. The sixth column reports the average and median across the 19 industries for the differences in the within-industry standard deviation between straight-line and present value amortization for capitalized operating leases, the frequency of positive and negative differences for the 19 industries, and the associated p-values. Probability values are based on the frequency of positive differences assuming a binomial distribution and a null hypothesis of 50 percent positive and 50 percent negative differences.

the 19 industries when replacing straight-line amortization with present value amortization, and this is significant at less than the 0.001 level for all three ratios.26 This indicates that comparability of financial statements is no worse when operating leases are amortized on a present value basis relative to a straight-line basis, and may be slightly improved. Thus, this analysis provides no 26

The within-industry standard deviation did not decline for present value amortization relative to straight-line amortization for asset turnover in the commercial and professional services industry (increase from 1.480 to 1.626), and for financial leverage in the retailing industry (increase from 0.298 to 0.346).



69

evidence in support of requiring straight-line amortization as the default amortization method for operating leases.27 CONCLUSION This study provides evidence on alternative methods for capitalizing operating leases, an issue that is currently being considered by the Financial Accounting Standards Board and the International Accounting Standards Board. First, we use a stylized example to compare the information provided by straight-line amortization of lease assets with the information provided by present value amortization of lease assets. Under ‘‘ideal’’ conditions, we find that present value amortization provides information that is more comparable across firms and that more faithfully represents the future cash flows of lease assets than straight-line amortization. We also find that both methods provide less faithful representations of future cash flows when lease assets are positive net present value investments. Second, using security prices as a reference point, we find that lease assets are viewed by investors as positive net present value investments for firms with above-industry-median profitability, but not for firms with below-industry-median profitability. We also find weak evidence that for firms whose profit margin is below the industry median, present value amortization is more consistent with how investors implicitly value the lease assets. Thus, this analysis also fails to provide justification for favoring straight-line amortization over present value amortization. Third, we use the distributions of three common financial ratios to examine the relative comparability of as-reported assets and liabilities with assets and liabilities resulting from capitalizing operating leases and amortizing them using either straight-line or present value amortization. We find that comparability is improved when capitalizing operating leases and amortizing them using straight-line amortization, and comparability is further improved by replacing straight-line amortization with present value amortization. Thus, consistent with the previous analysis, this analysis also fails to support the choice of straight-line amortization over present value amortization. In summary, our analysis provides no support for straight-line amortization as the de facto preferred method for amortizing right-of-use lease assets. In contrast, we provide some evidence that presents value amortization for right-of-use lease assets is consistent with investor valuations for many firms, and modestly enhances within-industry financial statement comparability for almost all industries.

REFERENCES Anton, H. 1956. Depreciation, cost allocation, and investment decisions. Accounting Research 7: 117–134. Ayers, B. 1998. Deferred tax accounting under SFAS No. 109: An empirical investigation of its incremental value-relevance relative to APB No. 11. The Accounting Review 85: 791–815. Barth, M. 1991. Relative measurement errors among alternative pension asset and liability measures. The Accounting Review 66: 433–463. Barth, M. 1994. Fair value accounting: Evidence from investment securities and the market valuation of banks. The Accounting Review 69: 1–25.

27

As with our previous analyses, while we are unable to explicitly test amortization based on the underlying asset approach, results for that approach would be similar to straight-line ( present value) amortization to the extent that the portion of the underlying asset consumed during the lease period is high (low).


70


Beattie, V., K. Edwards, and A. Goodacre. 1998. The impact of constructive operating lease capitalization on key accounting ratios. Accounting and Business Research 28: 233–254. Bennett, B. K., and M. E. Bradbury. 2003. Capitalizing non-cancellable operating leases. Journal of International Financial Management and Accounting 14: 101–14. Bierman, H. 1961. Depreciable assets—Timing of expense recognition. The Accounting Review 36: 613– 618. Boatsman, J., and X. Dong. 2011. Equity value implications of lease accounting. Accounting Horizons 25: 1–16. Bodenhorn, D. 1961. An economist looks at industrial accounting and depreciation. The Accounting Review 36: 583–588. Boone, J. 2002. Revisiting the reportedly weak value relevance of oil and gas asset present values: The roles of measurement error, model misspecification, and time-period idiosyncrasy. The Accounting Review 77: 73–106. Choi, B., D. Collins, and W. Johnson. 1997. Valuation implications of reliability differences: The case of non-pension post-retirement obligations. The Accounting Review 72: 351–383. Coughlan, J. 1959. Industrial accounting. The Accounting Review 34: 415–428. Davis-Friday, P., L. Folami, C-S. Liu, and H. Mittelstaedt. 1999. The value relevance of financial statement recognition vs. disclosure: Evidence from SFAS No. 106. The Accounting Review 74: 403–423. De Franco, G., M. Wong, and Y. Zhou. 2011. Accounting adjustments and the valuation of financial statement note information in 10-K filings. The Accounting Review 86 (5): 1577–1604. Dhaliwal, D., H. S. Lee, and M. Neamtiu. 2011. The impact of operating leases on firm financial and operating risk. Journal of Accounting, Auditing and Finance 26: 151–197. Dixon, R. L. 1960. Decreasing charge depreciation—A search for logic. The Accounting Review 35: 590– 597. Durocher, S. 2008. Canadian evidence on the constructive capitalization of operating leases. Accounting Perspectives 7: 227–256. Easton, P., M. McAnally, P. Fairfield, and X-J. Zhang. 2009. Financial Statement Analysis and Valuation. First Edition. Westmont, IL: Cambridge Business Publishers. Eccher, E., K. Ramesh, and S. Thiagarajan. 1996. Fair value disclosures by bank holding companies. Journal of Accounting and Economics 22: 79–117. Ely, K. 1995. Operating lease accounting and the market’s assessment of equity risk. Journal of Accounting Research (Autumn): 397–415. Feltham, G., and J. Ohlson. 1996. Uncertainty resolution and the theory of depreciation measurement. Journal of Accounting Research 34 (2): 209–234. Financial Accounting Standards Board (FASB). 2009a. Discussion Paper Leases: Preliminary Views. Norwalk, CT: Financial Accounting Standards Board. Financial Accounting Standards Board (FASB). 2009b. Discussion Paper Leases: Comment Letter Summary. Norwalk, CT: Financial Accounting Standards Board. Financial Accounting Standards Board (FASB). 2010. Exposure Draft, Proposed Accounting Standards Update, Leases (Topic 840). Norwalk, CT: Financial Accounting Standards Board. Financial Accounting Standards Board (FASB). 2011. Exposure Draft, Proposed Accounting Standards Update, Leases (Topic 840: Comment Letter Summary). Norwalk, CT: Financial Accounting Standards Board. Financial Accounting Standards Board (FASB). 2012. Summary of Board Decisions (February 27). FASB/ IASB Joint Board Meeting. Norwalk, CT: Financial Accounting Standards Board. Hotelling, H. 1925. A general mathematical theory of depreciation. Journal of the American Statistical Association 20 (151): 340–353. Imhoff, E., R. Lipe, and D. Wright. 1991. Operating leases: Impact of constructive capitalization. Accounting Horizons 5: 51–63. Imhoff, E., R. Lipe, and D. Wright. 1993. The effects of recognition versus disclosure on shareholder risk and executive compensation. Journal of Accounting Auditing and Finance 8 (4): 335–368. Accounting Horizons March 2013


71

Imhoff, E., R. Lipe, and D. Wright. 1997. Operating leases: Income effects of constructive capitalization. Accounting Horizons 11 (2): 12–32. International Accounting Standards Board (IASB). 2012a. Staff Paper, IASB Agenda Reference Addendum to 2C, Lessee Accounting—Approaches (addendum regarding Approach C). (February). Available at: http://www.ifrs.org/CurrentþProjects/IASBþProjects/Leases/MeetingþSummariesþandþObserver þNotes/IASBþFASBþFebruaryþ2012.htm International Accounting Standards Board (IASB). 2012b. Staff Paper, IASB Agenda Reference 2D, Lessee Accounting—Illustrations. (February). Available at: http://www.ifrs.org/CurrentþProjects/ IASBþProjects/Leases/MeetingþSummariesþandþObserverþNotes/IASBþFASBþFebruaryþ2012. htm Jennings, R., P. Simko, and R. Thomson II. 1996. Does LIFO inventory accounting improve the income statement at the expense of the balance sheet? Journal of Accounting Research 34: 85–109. Nelson, K. 1996. Fair value accounting for commercial banks: An empirical analysis of SFAS No. 107. The Accounting Review 71: 161–182. Revsine, L., D. Collins, W. Johnson, and H. Mittelstaedt. 2009. Financial Reporting and Analysis. Fourth Edition. New York, NY: McGraw-Hill Irwin. Reynolds, I. N. 1961. Selecting a proper depreciation method. The Accounting Review 36: 239–248. Staubus, G. J. 1962. Decreasing charge depreciation—Still searching for logic. The Accounting Review 37: 497–501. Stickney, C., P. Brown, and J. Wahlen. 2007. Financial Reporting, Financial Statement Analysis, and Valuation. Sixth Edition. Mason, OH: Thompson South-Western. White, G., A. Sondhi, and D. Fried. 2003. The Analysis and Use of Financial Statements. Third Edition. New York, NY: John Wiley and Sons. Wild, J., K. Subramanyam, and R. Halsey. 2007. Financial Statement Analysis. Ninth Edition. New York, NY: McGraw-Hill Irwin.

APPENDIX A LEASE ASSET AMORTIZATION USING THE UNDERLYING ASSET APPROACH This appendix extends our example in Table 1 and the main text to apply the Underlying Asset Approach (UAA) amortization method, an alternative method for amortizing right-of-use lease assets under the new leasing standard proposed by the Financial Accounting Standards Board and the International Accounting Standards Board. At their joint meeting on February 28–29, 2012, the two Boards discussed this method as an alternative to the straight-line and present value alternatives already under consideration. UAA amortization views the lessee’s amortization expense from the perspective of the compensation required by the lessor for the lessee’s use of the underlying asset during the lease period. From this perspective, the lessee’s right-of-use asset amortization has two components. The first component is compensation to the lessor for consuming part of the underlying asset, and is computed as the straight-line amortization of the portion of the lessor’s underlying asset value that is consumed during the lease period. This component is referred to as ‘‘consumption of asset.’’ The second component is compensation to the lessor for the opportunity cost of being unable to use the unconsumed portion of the lease asset during the lease period. This is computed as the increase in value each year of the lessor’s end-of-lease residual value in the underlying asset, and is referred to as ‘‘residual accretion.’’ Table 7 presents results from applying UAA amortization to the example used in Table 1 and the main text under different scenarios. In each scenario, we assume a ten-year lease with annual payments of $1,000 at the end of each period, and an incremental borrowing rate for the lessee of 7 percent, so that the present value of the lease liability and asset at inception of the lease is $7,024. To apply the underlying asset approach requires an estimate of the portion of the underlying asset Accounting Horizons March 2013


72

TABLE 7 Example of the Underlying Asset Approach Amortization Method Panel A: Lease is Years 41–50 of 50-Year Asset Life Year of Lease 1

3

5

7

9

Cash Inflow Consumption of Asset Residual Accretion Income before Interest

1,000 702 0 298

1,000 702 0 298

1,000 702 0 298

1,000 702 0 298

1,000 702 0 298

Beginning Book Value SL Amort Rate of Return

7,024 4.2%

5,619 5.3%

4,214 7.1%

2,809 10.6%

1,404 21.2%

7

9

Panel B: Lease is Years 31–40 of 50-Year Asset Life Year of Lease 1

3

5


1,000 357 250 393

1,000 357 286 357

1,000 357 328 315

1,000 357 375 268

1,000 357 429 214


7,024 5.6%

5,792 6.2%

4,486 7.0

3,093 8.7%

1,603 13.3%

Panel C: Lease is Years 110 of 50-Year Asset Life Year of Lease 1

3

5

7

9


1,000 47 474 479

1,000 47 543 410

1,000 47 622 331

1,000 47 712 241

1,000 47 815 138


7,024 6.8%

5,948 6.9%

4,730 7.0%

3,349 7.2%

1,781 7.8%

Panel D: Lease is for Land That Does Not Depreciate Year of Lease 1

3

5

7

9


1,000 0 508 492

1,000 0 582 418

1,000 0 666 334

1,000 0 763 237

1,000 0 5,730 127


7,024 7.0%

5,971 7.0%

4,767 7.0%

3,387 1,808 7.0% 7.0% (continued on next page)



73

TABLE 7 (continued)

Each panel of this table presents an example of the Underlying Asset Approach (UAA) method of amortizing right-ofuse lease assets. In all four examples, the lessee pays $1,000 at the end of each of ten years, and uses an incremental borrowing rate of 7 percent to compute the value of the lease asset and liability at inception of the lease, which are both equal to $7,024. The net cash inflows from the leased asset are expected to be equal to the lease payments in both timing and amount, so that the lease is a zero net present value investment. Consumption of Asset is the straight-line amortization of the portion of the underlying asset’s value consumed during the lease, and is equal to the difference between the asset’s value at inception of the lease and at termination of the lease, divided by the length of the lease. The asset’s value at inception and termination of the lease is determined by replacing the length of the lease in the calculation of the present value of the lease payments with the remaining life of the asset at the beginning and end of the lease, respectively. Residual Accretion is the product of the present value of the residual value of the underlying asset and the incremental borrowing rate used to discount the lease payments (7 percent in this example). In Panel A, the lease is for the last ten years of an asset with a total expected life of 50 years. The asset’s residual value at the end of the lease is zero, so that consumption of asset is equal to 7,024/10 ¼ 702 each year, and residual accretion is equal to (0/(1.07)n)(0.07) ¼ 0, where n is the number of periods remaining in the lease. In Panel B, the lease is for years 31–40 of an asset with a total expected life of 50 years. The asset value at the beginning (end) of the lease is the present value of a 1,000 annuity discounted at 7 percent for 20 (10) years, and is equal to 10,594 (7,024). Consumption of asset is equal to (10,594 7,024)/10 ¼ 357 each year. Residual accretion is equal to (7,024/ (1.07)n)(0.07), where n is the number of periods remaining in the lease, and is equal to 250 in the first year. In Panel C, the lease is for years 1–10 of an asset with a total expected life of 50 years. The asset value at the beginning (end) of the lease is the present value of a 1,000 annuity discounted at 7 percent for 50 (40) years and is equal to 13,801 (13,332). Consumption of asset is equal to (13,801 13,332)/10 ¼ 47 each year. Residual accretion is equal to (13,332/ (1.07)n)(0.07), where n is the number of periods remaining in the lease, and is equal to 474 in the first year. In Panel D, the lease is for land that does not depreciate, so that the asset value at the beginning (end) of the lease is the present value of a 1,000 perpetuity discounted at 7 percent, and is equal to 14,286 (14,286). Consumption of asset is equal to (14,286 14,286)/10 ¼ 0 each year. Residual accretion is equal to (14,286/(1.07)n)(0.07), where n is the number of periods remaining in the lease, and is equal to 508 in the first year.

that is consumed during the lease period. Throughout our example, we do this by assuming that the fair value of the underlying asset to the lessor is determined by extending the terms of the lease to the entire life of the asset. In the first three panels of Table 7, we assume the asset has a total life of 50 years and that the lessor values the asset at each point in time on the basis of leasing the asset at $1,000 per year for the remaining years of its life. We then vary which ten-year period during that 50-year life is the lease period for the example lease, and this varies the portion of the remaining life of the underlying asset that is consumed during the lease period. In the last panel of Table 7, we assume the lease is for land that does not have a finite life, so that the portion of the asset’s value consumed during the lease period is zero. In Panel A, we assume that this lease is for the last ten years of the life of an underlying asset’s 50-year life. Under these circumstances, the portion of the remaining underlying asset consumed by the lessee during the lease period is 100 percent and the lessor’s end-of-lease residual value in the underlying asset is zero, so that the consumption of the asset is equal to (7,024 0)/10 ¼ 702 each year and residual accretion is zero. Panel A of Table 7 reports rates of return for UAA amortization under these circumstances that are exactly the same as those reported in Panel A of Table 1 for straight-line amortization for the same lease. In Panel B, we assume that the lease is for years 31–40 of the underlying asset’s 50-year life. Under these circumstances, the value of the underlying asset at the beginning (end) of the lease is the present value of a 1,000 annuity discounted at 7 percent for 20 (ten) years and is equal to 10,594 (7,024). This results in consumption of the underlying asset that is equal to (10,594 7,024)/10 ¼ 357 each year, and residual accretion that is equal to (7,024/(1.07)n)(0.07), where n is the number of periods remaining in the lease, and is equal to 250 in the first year. The rate of return presented in Panel B for year 1 (9) for this scenario is larger (smaller) than in Panel A, indicating movement


74


away from the straight-line amortization results presented in Panel A of Table 1 and toward the present value amortization results presented in Panel B of Table 1. This movement is more evident in Panel C of Table 7, where we assume that the lease is for years 1–10 of the asset’s 50-year life. Under these circumstances, the value of the underlying asset at the beginning (end) of the lease is the present value of a 1,000 annuity discounted at 7 percent for 50 (40) years, and is equal to 13,801 (13,332). This results in consumption of the underlying asset equal to (13,801 13,332)/10 ¼ 47 each year, and residual accretion equal to (13,332/ (1.07)n)(0.07), which is 474 in the first year. The rate of return presented in Panel C for year 1 (9) for this scenario is much larger (smaller) than for straight-line amortization in Panel A, and is much closer to the results for present value amortization presented in Panel B of Table 1. Finally, in Panel D of Table 7, we assume that the lease is for land that does not have a finite life. Under these circumstances, the value of the underlying asset at the beginning and end of the lease is the present value of a 1,000 perpetuity discounted at 7 percent, and is equal to 14,286. Thus, consumption of the underlying asset is (14,286 14,286)/10 ¼ 0 each year, and residual accretion is (14,286/(1.07)n)(0.07), which is 508 in the first year. The rates of return presented in Panel D for this scenario in all years are identical to those presented in Panel B of Table 1 for present value amortization. In summary, the Underlying Asset Approach to amortization of the right-of-use lease asset is a weighted-average of straight-line amortization and present value amortization, and this combination is more (less) weighted toward straight-line amortization as the portion of the lessor’s underlying asset consumed during the lease period increases (declines).28

28

Although we do not report details to save space, we find the same relationship among the three alternative amortization methods when the lease is a positive net present value investment as in Panels C and D of Table 1.
