Return on Assets So Useful...and So Misused

Return on Assets So Useful...and So Misused

Journal Issue: 
Column Name: 
Journal Article: 

Perhaps no financial ratio is as frequently cited and misunderstood as return on assets (ROA). When properly defined and analyzed, ROA—which we define here as trailing four-quarter operating profit (earnings before interest expense and incomes taxes, or EBIT) divided by average total assets—is a key measure of relative operating performance. It can be employed as a historical time-series for a firm and its operating segments, or as a benchmark measure of operating performance for a group of competing companies. For distressed companies, ROA is a high-level diagnostic tool that helps identify areas of underperformance. For companies attempting to reorganize through the bankruptcy process, ROA can serve as a gauge of reasonableness of the debtor's business plan and its underlying assumptions, but presents some unique pitfalls as well.

Breaking Down ROA

ROA is in fact the product of two other ratios: total asset turnover and operating margin, calculated as follows:

(Four-quarter Sales/Avg. Total Assets) * (Four-quarter EBIT/Four-quarter Sales)

The first ratio, total asset turnover, is a measure of a firm's productivity, i.e., how much revenue does one dollar of assets generate? The second ratio, operating margin, is a measure of a firm's profitability, i.e., what percentage of these revenue dollars remains as operating profits? Putting the two ratios together—which can be done algebraically by canceling out the sales terms—provides an easily calculable, high-level summary measure of operating performance. EBIT is generally the preferred measure of operating profit provided that depreciation expense (which has already been deducted from EBIT) is a fair proxy of maintenance-type capital expenditures. When depreciation expense is not a close approximation of routine capital expenditures (e.g., store operating leases of retailers, where no depreciation expense is recognized but maintenance of the property is often the responsibility of the tenant), the analyst may prefer to estimate the portion of historical capital expenditures that were non-growth related outlays, and deduct this estimate from EBITDA (earnings before interest, taxes, depreciation and amortization).

The Common Misconception About ROA

Myriad textbooks and other financial publications improperly define ROA as net income (sometimes pre-tax income) divided by total assets. We disagree with this definition because a proper measurement of ROA should focus exclusively on operating returns, and therefore should be independent of the firm's capital structure—which net income clearly is not. Generally speaking, ROA is part of the family of income/investment measurements, and as such, the manner in which the firm chooses to finance the investment should be irrelevant to the calculation of the return on that investment. Net income is an inappropriate choice of numerator as a measurement of income or return on investment because it includes a deduction for interest expense, which is a function of the firm's financing decisions. The fallacious logic underlying the incorrect definition of ROA would produce different ROAs for two firms with identical asset bases and operating returns but dissimilar capital structures—solely as a result of their differences in net income caused by interest expense. This conclusion is erroneous because the firm's financing decisions are influencing the measurement of operating returns.

Net income is a return to equity-owners; it belongs entirely to the firm's shareholders. It is the appropriate numerator for a return on equity (ROE) calculation. Operating profit, or EBIT, belongs to all of the firm's suppliers of capital, both debt and equity, and is the commensurate measure of income or return on the firm's total assets. Comparisons between ROA and ROE indicate how effectively the firm is using financial leverage. For a debt-free firm, ROA (using fully-taxed EBIT as the numerator) will necessarily equal ROE. For a firm that employs debt, ROE will exceed ROA whenever the firm is able earn a return on borrowed capital that exceeds the explicit cost of such borrowing.

Reconciling ROA and ROIC

Closely related to traditional ROA analysis is the concept of return on invested capital (ROIC), an operating metric favored by advocates of economic value-added (EVA) methodologies. EVA proponents take ROA analysis one step further by quantifying the enterprise value created by the firm when it earns an operating return on capital (however this capital is defined) that exceeds its weighted average cost of capital (WACC). However, in contrast to the use of total assets as the denominator or investment base, EVA adherents would argue that invested capital is the appropriate denominator, defined as the book value of total debt plus shareholders' equity, or total assets minus current liabilities (excluding any current portion of debt). The logic behind this adjustment is that some portion of the firm's total asset base has been financed without cost to the firm by vendors, employees, taxing authorities, etc., and these amounts should be excluded from the asset base when computing operating returns. In other words, the investment base on which we measure the firm's operating returns should only include permanent capital committed to the firm. Furthermore, EVA calculates both ROIC and WACC on an after-tax basis, so EBIT or operating profit is usually adjusted and expressed as NOPAT, or normalized operating profit after taxes.

EVA "pioneers" of the late '80s advanced the popularity of that ubiquitous term "value creation," which is derived from the capitalized value of the annuity-like cash flows generated by the firm's operations. The firm creates value that accrues to its shareholders when it earns a return on invested capital that exceeds the cost of this capital. The value created (or destroyed) each year is equal to the positive (negative) difference between ROIC and WACC multiplied by average invested capital. If this value is created consistently year after year to the point of virtual permanence, then it can be capitalized just as any perpetuity can—by dividing it by an appropriate discount rate. The capitalized expected value of this perpetuity is the value premium of the firm's equity.

What Does ROA Tell Us?

Once ROA has been calculated, it is imperative to examine how its two components (total asset turnover and operating margin) influence its composition and its changes over time. What really drives a firm's ROA—productivity or profitability—and in what proportions? The topic is best illustrated by example.

Retailers of GAF category products (General Merchandise, Apparel and Accessories, Furniture and Home Furnishings) typically fall into one of three retail channels or store formats: department stores, discount stores or specialty stores. We ran a screen for publicly owned retailers with annual sales of at least $100 million and SIC codes that correspond to any of these three store formats. Eighty-one companies were identified. An analysis of ROA for these 81 retailers (unweighted) over the past 16 quarters (see Chart 1, with 0Q representing the most recent quarter of 2001) reveals that specialty stores have consistently outperformed the other two formats—with almost double the ROA of department stores in 1998 and 1999. What explains this disparity in operating performance? EBIT margins of specialty stores—the first component of ROA—have historically been less than 100 basis points higher than department stores (see Chart 2)—not nearly enough to explain the difference. Ultimately, it is superior store productivity as measured by total asset turnover—the second component of ROA—that drives the operating returns of these specialty retailers. Their business requires significantly smaller investments of capital relative to the sales they generate or, conversely stated, specialty retailers generate far more sales dollars for each dollar of assets. Chart 3 indicates that specialty retailers enjoy an asset-turnover ratio that is 50 percent higher than department stores; they generate about $2.25 of sales for every $1 of assets compared to $1.45 to $1.00 for department stores. Chart 3 also highlights that poor store productivity is endemic to traditional department stores, many of which have been the target of frequent criticism by retail analysts during the second-half of the '90s. They have been harshly characterized by some as lumbering, monolithic relics of a bygone shopping era that have become irrelevant to today's consumers, with the exception of well-heeled, middle-aged women. More to the point, the emergence of category killers and supercenters that began in the early '90s has caused many consumers to migrate away from traditional department stores for their non-apparel needs, thereby reducing the sales productivity of this venerable shopping venue.




Department and discount stores both generated poor returns on book assets in 2000, but the sources of underperformance were quite different. Department stores have been chronic ROA laggards due to low sales productivity arising directly from the costly structures in which they choose to operate. This is an embedded aspect of their business that will likely prove difficult to materially improve going forward. Conversely, discount stores have relied on exceptionally high sales productivity (Chart 3) to offset meager operating margins—a formula that worked pretty darn well in the '90s. However, discount stores took an unusual and severe ROA hit in 2000 (Chart 1) due to unanticipated profitability issues (Chart 2)—typically merchandising, pricing or inventory issues—which can be more quickly addressed and more easily remedied than the long-term, structural issues facing many department store chains today.

Adjusting ROA for Occupancy Costs

Occupancy costs present a particularly thorny problem for ROA comparisons whenever the accounting treatment of such costs differs among competitors and the financial effect of these accounting differences materially affects the ROA calculation. For example, certain old-line retailers, such as traditional department stores, capitalize a sizeable portion of their operating properties and related fixtures—which are either owned outright or, more frequently, financed with mortgages or capital leases. In these instances, an asset and corresponding obligation are recognized on the balance sheet, and monthly payments related to occupancy costs are divided between interest expense and obligation reduction, with no rent expense recorded. Alternatively, most specialty retailers occupy the vast majority of their store properties via operating leases that have no balance-sheet impact at all. Rent expense associated with operating leases is recorded, reducing operating income (EBIT). Conversely, interest expense associated with capital lease or mortgage payments is not a component of EBIT, but rather is shown as a separate line-item below operating income in the statement of operations.

Accordingly, adjustments must be made to the basic ROA calculation to ensure comparability across companies regardless of the accounting treatment of occupancy costs. First, future operating lease obligations—as disclosed in the footnotes to the annual financial statements—must be brought onto the balance sheet (as both an asset and a liability) so that all companies will have effectively capitalized all their "boxes" regardless of how these boxes are occupied. The capitalization of future operating lease obligations is most handily accomplished by dividing the current-year rent expense by a market-based cap rate. This capitalized amount should then be added to the denominator used in the calculation of ROA. Secondly, annual rent expense on operating leases should be added back to EBIT, the numerator of the ROA calculation. In the previous example, these adjustments would have reduced the ROA of specialty retailers by between 100-200 basis points annually, but the basic conclusions remain intact.

ROA and Fresh-start Reporting

ROA comparability across companies or across time may also be compromised when a company has emerged or will emerge from bankruptcy and adopts fresh-start reporting upon emergence in accordance with the provisions of SOP 90-7. Fresh-start reporting requires that the emerging debtor determine its reorganization value, which approximates the fair value of the entity before considering any existing liabilities. This fair value is generally determined by means of discounted cash flow (DCF) methodology—that is, by applying a market multiple to the projected operating income (EBIT) of the reorganized entity and then discounting this amount to present value. The reorganization value becomes the basis for the emerging entity's assets on its first post-emergence balance sheet, and is subsequently allocated to the specific assets of the reorganized company. Given the precarious prospects of most reorganized entities, conservative operating assumptions and multiples are typically applied in the determination of reorganization value. This tendency toward conservatism frequently results in a reorganization value that is significantly, perhaps drastically, below the debtor's pre-emergence asset values based on historical costs.

If these conditions prevail, then measurements of ROA for a reorganized entity after its emergence would be biased on the high side as a result of a denominator that has little, if any, relationship to either the historical costs or the replacement costs of the entity's actual operating assets. Consequently, ROA comparisons to non-bankrupt peers may leave the mistaken impression that the emerged entity is achieving acceptable—even superior—operating returns, when in fact adopting the fresh-start reporting standard by the reorganized entity rendered these comparisons utterly meaningless.

An extreme example of the distortion that can occur with fresh-start reporting is Ames Department Stores. Ames recently filed for bankruptcy again after first emerging from bankruptcy back in 1992. It adopted fresh-start reporting in December 1992, and valued the reorganized entity at $665 million (or 4.25 times Projected Year Four EBIT)—some $450 million less than the comparable total asset amount reflected on its final pre-emergence balance sheet. This valuation was so low that it was exceeded by the fair value of Ames's current assets, resulting not only in a write-down of all non-current assets (including all PP&E) to zero, but the rare recognition of negative goodwill, which it amortized into net income over subsequent periods. Calculations of ROA for Ames in subsequent periods would have been distorted in several ways: (1) EBIT included virtually no deduction for depreciation expense (our usual proxy for routine capital expenditures) since historical non-current assets had already been reduced to zero; (2) EBIT was increased by the amortization of negative goodwill—an accounting construct unrelated to operating performance; and (3) the investment base, or denominator, upon which operating returns are measured was wholly unrelated to the approximate economic investment in assets required to generate these returns.

Strictly using post-emergence accounting data would have overstated Ames's ROA, and comparisons to other discount retailers would have been misleadingly favorable. In these extreme instances, there are no simple reconciling adjustments that necessarily make ROA comparisons across companies meaningful. Clearly, Ames's pre-emergence historical asset values would have been an equally inappropriate choice as an investment base. The analyst must recognize the impact of fresh-start adjustments on the calculation of ROA and then determine whether the severity of these adjustments renders ROA useless as a measurement of relative economic returns.

User Beware

ROA remains a popular operating metric due largely to its intuitive appeal and relative ease of computation. Its primary limitations lie in its failure to account for the relative risk and size of companies. We previously established that the ROA of specialty stores consistently exceeded those of discount stores and department stores for the past four years, and well it should have! Specialty stores, with narrowly defined product categories, assume greater operating risk and are more likely to encounter distress in the event of a prolonged economic downturn. Discount stores and department stores are far more diversified in their product offerings—selling everything from soap to large appliances—and generally have better survival prospects under the same inclement business environment. Whether a specialty store generates an ROA large enough to compensate for its additional operating risk is a consideration ignored by ROA. This problem is mitigated when ROA is used as a comparative measure across similar competitors or to analyze trends over time for the company or its business units.

The importance of dollar returns relative to percent returns needs to also be considered when using ROA. An 8 percent ROA from a $50 billion retailer is probably more impressive than a 10 percent ROA from a competing $100 million retailer. One shouldn't jump to the conclusion that the larger ROA necessarily belongs to the "better" company. The analyst can mitigate this concern by being mindful of size considerations and differences when selecting a peer group of competitors and when making ROA comparisons.

Lastly, blind adherence to any principle or dogma is far easier on the brain than is scrutiny, and this applies even to the very secular realm of ratio analysis. In general, ratio analysis tends to be a fairly rigid exercise, with pat, pre-defined formulas that mostly fail to take into account the complexities of business and the nuances of accounting policies. In calculating financial ratios, there is an overwhelming temptation for the user to simply provide the required accounting input and run with the answer. For some general ratios, this tendency will not take the user too far away from the underlying truth. But for more focused ratios, such as ROA, blind adherence to a prescribed formula without some healthy degree of scrutiny can produce highly inaccurate results that lead the user to erroneous conclusions.

Bankruptcy Code: 
Journal Date: 
Thursday, November 1, 2001