Revenue is the top line of the income statement for a reason: It is the lifeblood of any company. Investors scrutinize trends in revenue growth to assess the strength of a company’s performance. But despite revenue’s crucial importance, they don’t question it nearly as much as other income statement items, such as earnings per share. The conventional wisdom is that revenue is harder to manipulate than EPS, leading some to include price/sales ratios in their analysis. Yet there are many ways management can mask a slowdown by manipulating the top line.
Any investor with a few years’ experience reading financial statements and a willingness to work can detect revenue manipulation. Yet most don’t, because reading footnotes seems like an arduous task. So management persists in massaging revenue in numerous ways to mask underlying weakness in a company’s reported results, because they can. Because the rest of the financials depend on revenue, it deserves the most scrutiny. Any doubt about the sustainability of revenue throws the rest of the financial model into question. Indeed, the financial fish rots from the revenue head.
Starting with the income statement, Figure 2.1 shows how to prioritize concerns from the top down.
Figure 2.1 Income statement: most critical to least critical earnings quality concerns.
Source: John Del Vecchio, © 2011.
It’s commonplace to hear value investors say that, of the three financial statements, the income statement is the least important and that the cash flow statement and balance sheet are paramount. But revenue is a key driver of all three statements.
Accurate revenue is crucial to confidence in cash flow, because the bottom line of the income statement—net income—is the top line of the cash flow statement. Investors read the cash flow statement in many ways, calculating Warren Buffett’s owner’s earnings; so-called “true” free cash flow, which includes changes in working capital; free cash flow to the firm; and more. But it is just as important to know whether revenue manipulation affects the quality and sustainability of cash flow.
Revenue affects the balance sheet, too. Consider deferred revenue, which can signal that demand for a company’s products is waning or that the company aggressively recognizes revenue. Revenue ties all three financial statements together. Revenue recognition is not the only element of earnings quality analysis, but it absolutely is the most critical.
Revenue is recognized when it is earned, but the timing is open to interpretation. The Financial Accounting Standards Board (FASB) has provided some concepts to guide management,1 but they leave enough discretion to lead to questionable practices and results.
Revenue is “realized” when goods, services, merchandise, or other assets are exchanged for cash or claims to cash. There is a large difference between a dollar deposited in the company checking account and having a claim on one. Not for nothing do they say “a bird in hand is worth two in the bush.” Claims to cash may require collection, and not everything may be collected. Revenues realized today according to this definition vary from certain to iffy.
“Realizable” is murky too. Revenues are “realizable” when the receivables are “readily convertible to known amounts of cash or claims to cash.” Instead of a claim to cash, the company can “readily convert” to “known amounts” of “claims to cash.” Management perceptions of types of claims may and do differ.
“Earned” revenues allow management the most latitude for questionable practices. Here, revenues are earned when the company “has substantially accomplished what it must do to be entitled to the benefits represented by the revenues.” It doesn’t take a lawyer to find room in “substantially,” “must,” “entitled,” and “benefits” for whatever timing the company wants.
With such gray definitions, management can wiggle, especially for intangibles, such as software. Evaluating whether management sticks to the straight and narrow or pushes the limits tells you how much confidence you can have in management and whether the company’s earnings quality is likely to be poor or good.
Here are some of the factors that we analyze to determine the quality of the company’s revenue. And at the top is a metric you should have burned into your memory: days sales outstanding.
A business wants to be paid as soon as possible. The longer it takes for customers to pay invoices, the more that company in essence provides no-interest customer financing—becoming a lender intentionally or not. Receivables can warn of a slowing economy, sure, but investors don’t own the economy. They own specific companies that may or may not track the economy, and if they do so negatively, may want to put off showing it. Tracking receivables is crucial, because trends may show, not only poorer customer financial condition, but also that the company may be changing customers’ payment terms to keep things looking good.
How many days it takes a company to be paid, on average across its customer base, is called days sales outstanding (DSO). Changes in DSO tell us a lot about whether the payment terms have shifted. More favorable payment terms often equal more aggressive revenue recognition. To calculate DSO2:
By tracking the receivables over several quarters—a moving aver-age—we get a better sense of the trend. We want to compare the quarter’s DSO both year-over-year (to account for business seasonality) as well as sequentially (from one quarter to the following). Changes can come from economic slowdowns, poor collection, and so on, but the practice we want to catch is a change in payment terms that borrows revenue from the future to make this quarter look better than it is.
Here is the key point that Wall Street gets wrong almost 100 percent of the time. Often, a company will report higher DSO, implying looser payment terms. Then management will comment in the quarterly conference call that the customers are of high quality, with good credit, so the company expects to collect on the receivables, and there is no risk to collection (la la la). Wall Street analysts usually take management’s word for it and then regurgitate this explanation in their own research reports.
This is absolutely wrong! Higher DSO has nothing to do with collection but everything to do with revenue recognition. If the company did not offer extended terms, it wouldn’t have been able to book the revenue in that particular quarter, regardless of whether it collects on the receivables. The company uses looser terms to pull revenue forward into the current quarter. This is called “stuffing the channel.” Here’s how it works.
A company may realize that the current quarter won’t meet Wall Street expectations, which may lead investors who are focused on quarterly results to sell the stock. Management might offer extended payment terms to induce customers to accelerate the purchase of goods and services so the company can book the revenue in the current period, boost results, and avoid a stock sell-off.
This actually happens a lot more than people think. Let’s say it’s June 28, with two days to go in the quarter and the company has booked only 80 percent of expected quarterly revenue. It’s in a bit of a jam. The CEO gets on a plane. He and his team sit across the table from you, a customer they know is all but certain to buy in the next quarter. That’s why they pick you. Under pressure, they don’t have time for a sales job, and they know you’re already sold. So they lay out that if you sign before the end of the quarter, the company will discount the price and offer more favorable payment terms. This happens all the time, especially in software, where customers know or learn to wait until the end of the quarter.
Of course, this only steals revenue from future quarters when the customer would buy anyway—and likely at more normal terms. This tactic can work if a company is growing quickly and the quarter is an anomaly, but more often than not, “stuffing the channel” only means there is less in the channel later. It works, until it doesn’t. The company becomes a mouse on a treadmill, and like a horror film, it ends badly.
The clearest—and rarest—indicator is if the company discloses in the Liquidity and Capital Resources section of SEC filings whether it offered special financing during the particular quarter. Good luck finding that, though. Companies that believe they have to stuff the channel to keep up appearances are hardly going to pop the balloon by revealing what they’re doing.
The alert investor who spies an increase in DSO knows that this can indicate special financing. Extended payment terms by definition mean an increase in DSO. Track days of sales outstanding over the past six to eight quarters to see if there are seasonal or other trends that make this quarter normal, or whether the increase in DSO means unusual activity in accounts receivable. Code words also can reveal a lot. If management alludes to a “heavily back-end-loaded quarter,” it may mean the company signed deals at generous terms near the end of the quarter in order to feed the Street. Raise your eyebrows if you hear “hockey stick growth” or “non-linear growth” for anything but a hypergrowth company. No company admits to stuffing the channel, so they use these other positive terms.
Here’s a telling anecdote. Tom purchased a low-cost cloud based software product from a fast-growing company. He used it for himself only—he was not a large business customer buying the software for 1,000 seats. Yet the salesperson kept calling daily the last week of the quarter. When Tom questioned, the salesperson said this was “normal” for the end of each quarter. Indeed, stuffing the channel is rife at software companies. Tom wished he had tried to get a better price, but he did move the salesperson off company insistence on a one-time annual payment in advance. He asked for periodic payments to smooth out cash flow. Done! If a company pressures its salespeople for a deal on this puppy, it’s probably offering far greater concessions to the big dogs.
A real-time case came in March 2006 when John issued this report to clients. He pointed to the risk increased DSOs posed to eResearch Technology investors:
eResearch Technology provides cardiac safety solutions to drug and medical device makers to evaluate new products. For the quarter ending Dec. 2005, DSO jumped on both a sequential and year-over-year basis. Table 2.1 shows DSO increasing approximately four days sequentially and five days year-over-year. The increase in DSO was largely driven by increased levels of receivables, despite lower revenue year over year.
Table 2.1 eResearch Technology DSO Issues*
In the quarterly conference call in February 2006, CEO Joseph Esposito noted that $8 million of business was signed in “the last days of December.” To the extent that certain of those revenues were booked rather than deferred, this may have indicated that the quarter was heavily loaded toward the end. And that pointed to possible aggressive terms or price discounts, either of which lowers the period’s quality of earnings.
It took only two months for the revenue to come home to roost. The arrow in Figure 2.2 shows where John issued this report on March 13. On April 26, the stock closed at $13.74, and the company announced lowered guidance for the rest of the year. Shares fell steadily, until they hit $9.12 on June 12, the day the company announced that Esposito would step down. The stock bottomed in the mid-$6s in November.3
Figure 2.2 eResearch: 2004–September 30, 2007.
Source: FreeStockCharts.com, used by permission.
Helen of Troy’s beauty products failed to obscure poor earnings quality when, in the first quarter of 2005, John alerted clients of concerns that the company was aggressively recognizing revenue.
Increase in A/R Days Due to Extended Payment Terms In the November 2004 period, the company experienced an increase in days sales outstanding compared with the year-ago period. At the end of the quarter, DSOs—according to the company’s calculation—were 72 days, an increase of three days, year over year. The company’s management attributed the increase to growth in international sales, which have longer credit terms on average. We are generally concerned about extended payment terms, as often they act to pull future revenue forward into the current period, leaving a revenue gap needing to be filled in an upcoming quarter. This increases the risk of an earnings miss in that time period, and as a result, may have an adverse impact on the stock price.
In John’s experience, even just a few days year-over-year increase in DSO can signal poor earnings quality and potential trouble ahead.
Percentage-of-completion accounting matches revenue earned on long-term contracts with the work performed by companies to fulfill their obligations. During each reporting period, management estimates the proportion of work completed and recognizes revenue and profits accordingly. Thus the company recognizes the revenue in advance of billing the customer.
The operative word is “estimates.” Any time discretion enters into the equation, there is the opportunity to manipulate the numbers. Management may underestimate costs incurred or overestimate the proportion completed in a given period, providing a boost to current results at the expense of future reporting periods.
Aggressive revenue recognition under the percentage-of-completion method may be determined a couple of ways. If management overestimated the proportion of work performed on the contract, a sharp increase in unbilled receivables relative to revenue should alert the investor. In addition, a sharp and unexplained increase in gross profit margin may indicate that the company’s management increased its estimated profits on the project.
Accelerated revenue recognition through percentage-of-completion accounting and other aggressive revenue concerns at AsiaInfo Holdings led John to release a report to clients of his Parabolix Research service on June 2, 2010. AsiaInfo shares stood at $21.54:
Company Description
AsiaInfo Holdings, Inc. provides telecommunications software solutions and information technology (IT) security products and services to telecommunications service providers, as well as to other major enterprises in China.
These two main factors lead to the conclusion that the company has deteriorating earnings quality: (1) accelerated revenue recognition and (2) likely overstated profit margins.
1. Trends in AsiaInfo’s Accounts Indicate Accelerated Revenue Recognition
AsiaInfo’s unbilled accounts receivables have surged in recent quarters, indicating possible revenue recognition acceleration. When a company uses percentage-of-completion accounting, it will have an unbilled receivable, which represents revenue recognized in advance of billing the customer. Aggressive managements may use the unbilled receivables as a way to front load revenue recognition from projects that are not yet completed. In essence, it is an estimated account. What’s more, management may estimate the profitability of the projects prior to completion as well. Table 2.2 illustrates AsiaInfo’s DSO trends.
Table 2.2 AsiaInfo Holdings’ Increasing DSO: June Quarter 2008– March Quarter 2010*
As Table 2.2 shows, the level of DSO has increased year-over-year for four consecutive periods. However, AsiaInfo’s accounts receivable levels require adjustments due to a relationship with IBM known as the “IBM Arrangement,” whereby the company acts as a distributor for certain products to China Mobile, its largest client. The language from the SEC filing follows:
In addition, in recent periods we have begun to generate service revenues by acting as a sales agent for International Business Machines Corporation, or IBM, or its distributors, for certain products sold to China Mobile, or the IBM Arrangement. The service fee under the IBM Arrangement is determined as a percentage of the gross contract amount. We have evaluated the criteria outlined in guidance issued by the Financial Accounting Standards Board, or the FASB, regarding reporting revenue gross as principal versus net as an agent, in determining whether to record as revenues the gross amount billed to China Mobile and related costs or the net amount earned after deducting hardware costs paid to the vendor, even though we bear inventory risks after the vendor ships the products to us and we bill gross amounts to China Mobile. We record the net amount earned after deducting hardware costs as agency service revenue because (1) the vendor is the primary obligor in these transactions, (2) we have no latitude in establishing the prices, (3) we are not involved in the determination of the product specifications, (4) we do not bear credit risk because we are contractually obligated to pay the vendor only when China Mobile pays us, and (5) we do not have the right to select suppliers.
Table 2.3 adjusts AsiaInfo’s DSO to reflect the IBM Arrangement. After adjusting for—netting—the IBM receivables, it appears as though virtually all [emphasis original] of AsiaInfo’s increase in accounts receivables is driven by changes in unbilled accounts receivable. Not only is the trend significant, but also so is the magnitude, compounded by the fact that ASIA has a heavy customer concentration with China Mobile representing 70 percent of AsiaInfo’s revenues, up from 66 percent last year and with the top three customers representing 93 percent of sales.
Table 2.3 AsiaInfo’s Increasing Unbilled Accounts Receivable, June Quater 2008-March Quater 2010*
While AsiaInfo’s unbilled A/R has been surging, deferred revenue has fallen sharply. The depletion of this account not only has served as a drain on cash flow, but also portends a slowdown in revenue in future quarters. Days in deferred revenue have dropped year-over-year for four consecutive periods. Table 2.4 shows that, as DSO have grown, the relationship between DSO to deferred revenue days (DDR) ballooned to 119 days in March 2010 from 57 days a year ago.
Table 2.4 AsiaInfo’s Worsening DSO Minus DDR: June Quater 2008-March Quater 2010*
2. Profit Margins May Be Overstated
While percentage-of-completion accounting can lead to accelerated revenue recognition by increasing revenues booked in advance of billing the customer, it may also overstate profit margins as management estimates the profitability of the projects in the interim. Again, it’s a management-estimated account.
Interestingly, while AsiaInfo’s unbilled A/R has surged, so has its profitability. Table 2.5 shows the recent trends in gross and operating profit margins. Gross margin exceeded 61 percent in the March 2010 quarter, a 750 bps improvement (bps = basis points; 100 basis points = 1 percent). Operating profit margin improvement topped 1,000 bps.
Table 2.5 AsiaInfo’s Improving Margins, June Quarter 2008-March Quarter 2010: Sustainable?
Profit margins were above expectations. When sell-side analysts questioned management about the margins in order to model their future estimates, replies were ambiguous at best. CEO Steve Zhang explained that the improvement in margins was due to several factors: “I think several reasons coming together drive our first quarter strong margin improvement. I think, first of all, the top line is driven by the—our Telecom Solution revenue. On the expense side, we did experience some seasonal delay in making new hires in the first quarter. Also, in the first quarter there was this Chinese New Year and everybody took a two-week vacation, so that also decreased our total expense. I think overall its top line [is] growing faster than the expense growth.”
One plausible explanation for higher top-line growth without corresponding expense is accelerated revenue recognition, as witnessed by the increase in unbilled A/R. The combination of surging unbilled receivables, declines in deferred revenues, and jump in profit margins significantly heightens the concerns with respect to AsiaInfo’s earnings quality.
The relationship among the three metrics suggests that any slowdown in revenue may have a material impact on AsiaInfo’s operations. Furthermore, on December 6, 2009, ASIA announced the acquisition of one of its competitors, Linkage Technologies (paying about $60 million in cash and issuing 28 million shares). The acquisition will only obscure AsiaInfo’s growth metrics and financial comparisons further.
The quarter after the report, the stock fell 26 percent on an earnings announcement. Within 14 months the stock was in single digits. Figure 2.3 shows the downfall.
Figure 2.3 AsiaInfo Holdings: April 2010 to December 9, 2011.
Source: FreeStockCharts.com, by permission.
The next concept is familiar to anyone who has ever borrowed from or lent money or property to a relative or friend. At the corporate level, it’s potential trouble.
When a company derives an increasing portion of its revenue from affiliated entities, the investor should be concerned. An outside investor cannot be sure whether the transaction is really arms-length between the parties. Be skeptical.
The notes to the financial statements disclose related-party revenue. Determine revenue without the related-party revenue and compare that to prior periods. If the revenue is not attractive without that from related parties, it’s a cause for concern. Sometimes related-party revenue is so large a part of total revenues that it’s a sure sign of trouble. UTStarcom (UTSI) showed this as the post–2000–2002 bull market gained momentum.
In 2004, approximately 80 percent of UTSI’s revenues came from China, and it sought to diversify, with a goal of 50 percent from outside China by 2006. The company noted its successful contracts with emerging markets in India, Vietnam, and Latin America as movement in the right direction.
But shareholders had no way to judge a large amount of the non-China revenue. At the time, 15 out of that 20 percent foreign revenue—that is, three quarters of it—was generated in Japan. The majority of that came from BB Technologies, a subsidiary of the Masayoshi Son-controlled investment bank, SoftBank. Masayoshi Son was chairman of the board of UTSI beginning in 1995. This alone doesn’t moot the astounding growth in UTSI revenues from BB Technologies from zero dollars in 2000 to $13.9 million in 2001 and $123 million in 2002—but surely Masayoshi Son’s connections were of some value in Japan’s interwoven corporate universe. And even a question about the quality of these revenues would discredit the vast majority of UTSI’s non-China income and its goal of achieving 50 percent from outside of China in a mere two years.
When the company set the goal in 2004 to diversity its revenue, UTSI was trading around $26. Within two months it had dropped 35 percent to $16.84, and two years later—the year it said it would reach the goal of having non-China revenues at 50 percent of total—it closed off 71 percent, at $7.50. And this was all during the roaring bull market. By July 14, 2011, shares had collapsed to $1.43. Investors alert to related-party income avoided this devastating loss or profited through shorting.
Investors had to be willing to comb the footnotes for questionable related-party transactions at Quest Software in 2005. John issued a report to clients warning that the notes disclosed major management and director conflicts of interests with shareholders. Quest acquired two companies. Quest’s Chairman and CEO and two other board members invested in a venture-capital fund that owned preferred stock in the two companies, entitling fund partners to cash payments of $95.6 million when the companies were acquired. Moreover, one of the board members was managing director of the fund. The clear conflict is that the fund’s interest was in selling at the highest possible price, while Quest shareholders’ interest was in paying the lowest.
While companies often have all sorts of cozy relationships among management, board members, and outside entities, rarely are they as blatant as at Quest. But they were blatant only to those who read the footnotes. And there management ended its discussion, concluding that its own interests were “not material.” Management is unlikely to call its own conflict “material.”4 It was an earnings quality analyst’s red flag.
The seemingly innocent “deferred revenue” line can also offer management all sorts of ways to make revenue appear better, but as in most of this chapter’s cases, it’s unsustainable.
Some companies, such as software firms, receive cash in advance of performing services for their customers. As a result, they defer this revenue—which means they delay recognizing it as revenue—until the services are performed. It’s crucial to watch the deferred revenue line.
Deferred revenue can be an indicator that revenue is rising or falling, especially when it is boosting reported revenue and may not be sustainable. The pressure to meet Wall Street’s quarterly expectations can push management to defer revenue. Of course, there is no bottomless well of deferred revenue from which to draw, so this practice just puts off, rather than eliminates, the inevitable quarterly earnings disappointment.
For example, if revenue grew $40 million sequentially, but deferred revenue declined by $20 million, management may have accelerated recognizing some or all of that $20 million to keep the growth rate up. Wall Street penalizes stocks when growth expectations aren’t met. But management can’t put off the real revenue picture forever. Without new deferred revenue recognition, which could be suspect as well, or new growth, it’s only a matter of time before a quarter brings real pain.
The next six places to look for trouble are not as predictive individually but together can further create a witches’ brew of a company’s aggressive revenue recognition.
Management can use nonmonetary transactions to boost reported revenue. For example, a company may provide products and services to a customer and in return receive stock, which may decline in value. It may also be illiquid, therefore tough to sell at any price. This was popular during the Internet boom as a way to show ever-increasing growth and fuel the stock price when companies paid each other in stock at inflated, unsustainable valuations.
John’s 2005 report flagged Helen of Troy’s non-monetary transactions in which the company exchanged its goods for advertising credits. The 10-Q filing describes the transactions:
During the fiscal year ended February 28, 2003, we entered into nonmonetary transactions in which we exchanged inventory with a net book value of approximately $3,100,000 for advertising credits. As a result of these transactions, we recorded both sales and cost of goods sold equal to the inventory’s book value. We had used approximately $2,000,000 of the credits through the fiscal year ending February 29, 2004. In addition, during the quarters ended August 31, 2004 and November 30, 2004, we entered into additional non-monetary transactions in which we exchanged inventory with a book value of approximately $952,000 and $59,000 respectively, for additional advertising credits. As a result of these transactions, we recorded both sales and cost of goods sold equal to the inventory’s book value, which approximated their fair value.
During the three months ended November 30, 2004, we used $368,000 of credits against these transactions and expect to use most of the remaining advertising credits acquired by the end of fiscal year 2005. All remaining credits are included in the line item entitled “Prepaid expenses” on our consolidated condensed balance sheets and “are valued at $1,743,000 and $1,100,000 at November 30, 2004 and February 29, 2004, respectively.”
These transactions do not appear to have any favorable impact on earnings because the book values of the products were recorded both as sales and cost of goods sold. However, we regard nonmonetary revenue to be of low quality, and it should be factored out of the analysis when assessing the true demand for a firm’s products.
While the “barter” here exchanged items of value, Helen of Troy could not use advertising credits to pay suppliers or employees. Those credits were only worth real sales if the advertising produced them. Such real sales from advertising were far from certain and the revenues, and therefore earnings, of low quality.
In this sleight of hand, the vendor invoices the customer and recognizes the transaction as revenue, but the products aren’t shipped until later. If this is a company’s practice, it should note the policies in its SEC filings.
Tom encountered this at an absurd level as a junior high–age investor. There actually was a company that recognized revenue of a finished product when it rolled off the assembly line and into storage. No customer, no invoice. This is the kind of fraud that shareholders simply can’t see. Tom invested $100 in the company because his best friend’s father served as general counsel. A go-go IPO during the late 1960s boom, the company said it had a modular housing construction technique that was going to answer the country’s affordable-housing problems. But when a housing unit came off the assembly line and left as inventory, the company didn’t even invoice a customer—it booked revenue immediately. The only way to know this was to suspect the rapidly rising revenue combined with equally fast-rising deferred revenue. Not only did the company eventually fail, because the revenue wasn’t real, it did so spectacularly enough to become a Harvard business school case study.5
A company may include in revenue an item that should be reported below the operating line. For example, from the 1990s to early 2000s, consulting firm Computer Associates included interest income as revenue. Although disclosed in footnotes, this would escape all but the most careful observer, who would see interest income in gross and operating margins and—ironically and misleadingly—in EBITDA. But interest income is not revenue because Computer Associates was not a bank, so the inflated sales figures made everything on down the income statement questionable at best. In November 2006, CEO Sanjay Kumar was sentenced to twelve years in prison for inflating sales figures in 1999 and 2000 and for other accounting misdeeds.6
A company can lease a product but treat it as a sale, a sales-type lease. Consider a copier company that leases to a business for a certain term. The company is allowed to recognize revenue up front by discounting future lease revenues to present value, but it’s aggressive. If using the operating lease method, the company is required to recognize revenue ratably over the term of the lease.
There is nothing inherently wrong with sales-type leases, but rapid growth in them can be a problem if accrual earnings outpace the growth in cash flows. All this does is bring future revenues forward, which can boost short-term results, but just puts off the day of reckoning.
When a company (buyer) purchases a distributor, it can slip by some double counting. The buyer has already sold items to the distributor and booked revenue. What the distributor hasn’t sold remains in inventory, which the buyer can sell and count again. Not only that, but the buyer can sell at retail the same inventory for which the distributor paid wholesale prices. An example is luxury accessory maker Coach’s 2005 purchase of the remaining 50 percent of its Japanese distributor, allowing Coach to post Japanese sales at the higher retail margin (what the Japanese distributor earned) than at wholesale (what the distributor paid Coach).
Change in revenue recognition policy should raise eyebrows, if the policy is changed to an earlier one of more aggressive revenue recognition. But a change to a more conservative policy can also boost results by double counting revenue. Management will take a charge to write off all prior revenue that would not have been recognized under the new revenue recognition policy. Instead, that revenue is deferred and re-recognized at a later date. The investor should compare revenue recognition policies over the past several periods and pay attention to changes in the policy language.
The investor has to get up pretty early in the morning—and wear magnifying glasses—to catch the true revenue effect of some changes in revenue policy. In June 2003, John notified clients that wireless intellectual property company InterDigital Communications had changed its policy to one that appeared to be more conservative—and would have been in many cases—but allowed the company to aggressively recognize revenue in this circumstance.
InterDigital made its money by licensing its patented technology to wireless companies, and that revenue stream was volatile and unpredictable. Figure 2.4 shows that license revenue declined in 1999 and again when the telecom bubble burst in 2000. After an almost flat 2001, revenue jumped 67 percent. But John advised skepticism about whether that growth could continue. This was in large part because InterDigital recognized revenue under one policy and then wrote it off when it changed to an allegedly more conservative policy. But Wall Street sees write-offs as one-time events and forgets about them. When the company recognized revenue under the new, allegedly more conservative policy, it appeared that revenues had risen. Of course, that too presented a problem, because the new recognition was one-time and the apparent growth was unsustainable.
Figure 2.4 InterDigital: 2002–October 31, 2003.
Source: FreeStockCharts.com, by permission.
Indeed, things did not play out well for InterDigital. During the wireless communications industry expansion in the early 2000s, companies such as InterDigital and Qualcomm held patents that they believed covered emerging technologies used in phones. Investors expected large paydays, believing that wireless phone companies would begin paying license fees and royalties.
InterDigital shareholders thought they saw the pot of gold when Nokia, which had fought like many phone makers against the patent holders, took a large reserve against potential liability for these license fees—and then raised that reserve. When the Finnish then-giant later changed its tack on paying the license fees to InterDigital, the latter’s stock tanked, as Figure 2.4 shows. The arrow indicates the June 13, 2003 issuance of John’s report.
A decade later, the FASB issued published revenue recognition rules for multi-deliverable contracts.7 The rules gave companies such as networking infrastructure maker Juniper Networks the opportunity to massage contracts for its benefit. In January 2011, John stated on CNBC that networking-equipment maker Juniper “may cough up a hairball in future quarters.” Indeed it did. John wrote at the time8:
Juniper Networks benefits from the Accounting Standards update at the beginning of fiscal 2010 that changed revenue recognition for multi-deliverable contracts, which Juniper adopted for all transactions after January 1, 2010.
Juniper sells software, hardware, and services to the communications industry, so contracts typically have multiple deliverables. Rather than defer the value of a contract and recognize the associated revenue ratably, the change allows companies like Juniper to break up the components of the contract and assign values to those components, recognizing some up-front, higher-margin components such as software. This gives the company enormous discretion on what to recognize and when, and the potential to front-load revenue. Year-over-year comparisons become meaningless.
Because of this change, effective January 1, 2010, the company recognized $237 million in revenue for the year. During the first three quarters after the change, it recognized $128 million, with the lowest amount—$23 million—in the quarter ending March 2010. The change’s anniversary comes in the first quarter of 2011, and the headwind to growth escalates as the year wears on. The revenue recognized in the fourth quarter of 2010 ($109 million) was more than double any prior quarter.
As a result, 26 percent reported growth drops to approximately 15 percent, when accounting for the change. This is approximately 5 percentage points (500 bps) below management’s forward guidance. To further grasp the magnitude for Juniper, the accounting change impacts competitor Cisco’s revenues by less than 1 percent.
The company also would have missed earnings estimates by approximately 10 percent a quarter over the last few quarters.
Since Juniper made the change, three metrics show the trouble. First, days in deferred revenue (DDR) have dropped each successive quarter to the lowest level since December 2008. The spread between the EBITDA Margin LTM and Operating Cash Flow margin LTM expanded to 300 bps from –400 bps in the year-ago quarter, suggesting that earnings quality has deteriorated progressively throughout 2010. [Chapter 5, “Cash Flow Warnings,” explains how to calculate and use these metrics.] An expanding spread indicated declining earnings quality. Finally, days sales outstanding have increased year-over-year the prior two quarters, acting as a drag on cash flow.
Juniper traded around $39 on the anniversary of the accounting change in April 2011 (see Figure 2.5). Six days later, the company reduced guidance and began to underperform the S&P 500. Investors may not have seen a one-time guidance reduction to be serious if they were in love with the stock, but when Juniper guided lower again on July 26, the affair ended. Shares on August 4 were $23, off 40 percent. Figure 2.5 shows the damage.
Figure 2.5 Juniper Networks: November 1, 2010–December 9, 2011.
Source: FreeStockCharts.com, by permission.
John alerted clients in 2005 that one-time events obscured real revenue growth at Helen of Troy. Due to recent acquisitions and foreign currency gains, Helen of Troy’s revenue growth was much weaker than reported. On a reported basis, revenue grew 24.4 percent, or $41.3 million, in November 2004, compared with the prior year period. However, new product acquisitions accounted for 21.1 percent or $34.9 million of that growth. Because of purchase acquisitions accounting, growth rates are inflated, simply because prior year periods are not restated to reflect the impact of the acquisition on historical results.
With respect to foreign currency, the strengthening of the British Pound, Canadian Dollar, and the Euro relative to the U.S. Dollar positively affected reported revenue. According to the company’s 10–Q filing for the November 2004 quarter, the net impact of foreign currency changes was to provide approximately $2.5 million of additional revenue relative to the year-ago period. While we will not speculate on the value of the dollar relative to foreign currencies, we do believe this to be a low quality source of revenue, due to its unsustainable nature.
Table 2.6 illustrates adjusted revenue to account for both the impact of acquisitions and foreign currency gains. An adjustment to reflect the impact of these other items on revenue lowers the company’s growth rate from the 24.4 percent reported amount to just 1.8 percent for the November 2004 quarter.
Table 2.6 Helen of Troy’s New-Product Acquisitions and Currency Gains Overstate Revenues*
John’s report concluded:
In our opinion, Helen of Troy’s earnings quality is low. Extended payment terms combined with a roll-up strategy, a weak U.S. dollar, and nonmonetary transactions pose several revenue-related concerns. Furthermore, the company’s cash flow performance has been underwhelming, as free cash is negative and operating cash flows are weak relative to reported net income. Finally, financial leverage has increased substantially due to recent acquisitions; however, its return on equity has barely budged.
The black arrow in Figure 2.6 shows the price on the date of the report. Helen of Troy was the face that launched a thousand slips.
Figure 2.6 Helen of Troy: 1996–April 30, 2009.
Source: FreeStockCharts.com.
Of all the financial statement red flags for companies, revenue recognition is the most serious and least studied. Aggressive revenue recognition is at the heart of trouble across all the financials. Next in importance is aggressive inventory management, which also flows through all the statements and offers companies many ways to obscure true earnings quality.
