Having discussed our strategy and implementation in the abstract, we thought we would give you a few real world examples to see how it can play out in practice. We chose the first two cases to illustrate how Wall
Street can ignore a large number of what were, to us, fairly obvious red flags. The third can then serve as an illustration of how the existence of, what we think are, red flags need not spell trouble (for the owners of the stock that is).
Before moving on, we offer a short apology. Although it might make for more interesting reading, we have withheld the names of the stocks in question for our own protection. The management teams of compa- nies that we short have lots of weapons at their disposal (if they choose to use them) to fight short sellers, including initiating short squeezes or forcing us to spend large amounts of time and money fighting legal bat- tles. We would rather focus our time on research.
Example 1: The Information Technology (IT) Outsourcer
This IT Outsourcer showed up on our screens at first because it was reporting strong net income and EPS growth while cash flow was actu- ally declining. This was particularly interesting because the company signed multiyear outsourcing contracts with its customers and recog- nized revenue on these under the percentage-of-completion (POC) method. Under POC accounting, a company recognizes revenue during a given period equal to the total revenue for the contract’s duration mul- tiplied by the ratio of costs incurred during the period to total estimated costs for the entire contract life. (That is, the company estimates a total profit margin for the contract and recognizes it as costs are incurred.) The POC method is a perfectly reasonable and generally appropriate for long-term contracts. Under POC, a company can overestimate contract profitability if it underestimates total costs. It can thus report higher lev- els of profitability in the early parts of a contract, and announce massive losses near completion.10 For the outsourcer, however, cash flow is reflective of the underlying economics of the business, and is thus not subject to estimation. While cash flow may fluctuate during any given quarter, if there is a recurring disconnect between it and earnings, one should ask why. We saw a number of red flags here.
For one, the company’s cash flow and EPS were moving in different directions. For its then most recently reported quarter, EPS had climbed by 14% from a year earlier while CFFO fell 6%. This drop in CFFO stemmed in part from rising DSOs, which had risen steadily over the pervious three quarters. Quarterly depreciation and amortization expense (D&A) as percent of PPE and intangibles fell to 4.6% during the quarter compared to 6.2% the previous one and 6.7% a year earlier.
10In fact, some defense companies with a conservative bias tend to have much lower profits at the beginning of a long-term complex contract and higher ones at the end, as some of their worst-case fears fail to pan out.
The ratio had not been below 6.2% over the last two years and had been as high as 8.5%. We thought that the company might have been stretching useful life assumptions or writing off assets to lower D&A. If the company had continued to report this expense at the same rate as it had in the previous quarter, it would have knocked $0.09 off the reported pro forma EPS of $0.69 (a figure a penny better than consensus expectations).
Meanwhile, SG&A expenses had dropped relative to revenue, giving one the impression that the company was doing an excellent job of con- trolling costs. Yet, as this was accompanied by significant growth in intangible assets, it occurred to us that costs might have been capitalized as part of some recent acquisitions. Contract signings, indicating future revenue trends, were increasing, but the rate of growth was declining.
More importantly, after adjusting for renewals, new signings declined 35% over the previous year. Furthermore, deferred revenue, represent- ing customers’ prepayments for future services and thus a good indica- tor of future sales, eased to 9.5% of revenue, from 10.8% in the preceding quarter and 10.6% a year earlier.
Looking further back, we found that in the prior 10 years, the com- pany generated cumulative free cash flow (CFFO less capital spending and related items) of just 2% of its current enterprise value (equal to market capitalization plus net debt). These weak cash flows helped lead to a $5 billion increase in the company’s total commitments (debt, pre- ferred stock, leases, minimum software subscription commitments, and other contingencies) in just three years, while the company posted prof- its of less than $3 billion over that time. Meanwhile, the company had spent $2 billion over the previous six months on four acquisitions that seemed only loosely related to its core business. Finally, several key executives had left or were leaving the company and a number of the remaining insiders had been selling shares.
Wall Street, however, saw a different picture. The charismatic CEO told a credible, conceptual, growth story based on long-term outsourc- ing trends. More importantly, the company almost always beat EPS expectations. The CEO really did deliver, saving many investors and sell-side analysts a lot of anxiety. Signings were strong, as was revenue.
Margins were expanding and EPS was steadily marching upward. What more could one ask for?
However, we thought that the cheerleaders might have been missing something. The company was operating in an intensely competitive industry, signing contracts that may have had very poor underlying eco- nomics. The business model had the company acquiring equipment and personnel from its customers while paying them cash upfront. It then hoped to recoup those investments over time. In order to entice custom-
ers to sign these contracts, it had to promise them savings by agreeing to receive fixed payments lower than the customer’s cost of running IT in- house. The company seemed to lack any clear competitive advantage over its peers, so that winning customer contracts may have been simply driven by price. All in all, we thought that this was probably a risky, low return business.
To make things worse, the IT industry was slowing down, yet the company seemed bent on reporting strong signing and EPS growth. If aggressive pricing could drive new business, then the earnings trajectory could have easily been maintained under POC accounting by bidding low for new contracts as long as the company “estimated” that the new busi- ness was profitable. Furthermore, management could have been seeking to acquire their way out of trouble by using a high stock price as currency.
We decided to “stress-test” our findings and set up a meeting with a sell-side analyst, who was one of the company’s leading champions on Wall Street. We presented our concerns and listened as he explained each away. We concluded that his explanations were indeed plausible, and frankly even better than ours for every single one. However, he could not address the fact that we had found so many of them at the same time. The probability that all were unrelated seemed low to us.
The analyst didn’t see our point.
We saw enough here to think that we may have been onto some- thing since there were no reasonable explanations we could uncover for the existence of so many warning signs. Meanwhile, the company’s stock was trading at 20 times forward EPS estimates that we thought were unattainable. We figured that this was a bet worth putting on.
Three quarters after putting on the position, the company’s shares came under pressure. Following several articles in the press questioning its accounting (which also might have provoked some additional ques- tions from the analyst community), the company made a series of announcements regarding problem contracts and delays. And it termi- nated negotiations with a few potential customers as well. Another two quarters went by and the company finally caved in and announced that EPS would be one fifth of original guidance. It is possible that the increased scrutiny limited its ability to manage earnings, or perhaps it ran out of levers to pull. Or just maybe the business deteriorated so much that it could not hide it anymore. In any case, the stock lost over 50% of its value following the EPS miss and over 70% compared to our entry point. We have to admit that we were not lucky enough to enjoy the entire ride. We did, however, reap a handsome reward.
It wasn’t that we were smarter than anyone on this one (we rarely are). We just have a hard time falling in love with a company and, hence, we let ourselves get bothered with the details.
Example 2: The Book Publisher
The Book Publisher initially showed up on our screens when it reported extremely strong earnings growth in its latest reported quarter while CFFO was actually declining. Furthermore, bad debt expense was down significantly over the past year. We decided to take a closer look.
First, while pro forma EPS had more than tripled from a year ear- lier, CFFO had actually dropped 35%. In addition, bad debt expense had fallen to 2.5% of revenue in the quarter from 4.6%, which could have boosted earnings by $0.10 per share (a material change as the com- pany reported $0.31 per share in the quarter, just meeting analysts’
expectations). Furthermore, the company took a massive charge three quarters earlier, leading to a significant drop in the amortization of pre- publication costs (advances, art, prepress, and other costs incurred in the creation of a master copy of a book). This deferral of some very real costs could have helped earnings by $0.07 per share. Furthermore, accrued liabilities, which may have contained reserves and other allow- ances, dropped. This may have been related to management under- reserving for bad debts or reversing reserves.
We figured that the publisher was operating in a no-to-low growth segment of the industry. However, it was fortunate enough to have one hot author under contract. She had released her first book in the previ- ous year, a blockbuster selling significantly more than expected. Her widely anticipated next book was due to come out about 12 months later and management established very high expectations for this release. In the interim, in an apparent effort to maintain earnings momentum, the company promised to cut costs.
However, it appeared to us that there was no real meat in the cost cuts. Of the $35 million supposedly achieved at that point, $6 million came from cutting allowances for doubtful accounts, $10 million was due to the fact that goodwill no longer needed to be amortized (as accounting standards had changed), and $17 million was due to lower prepublication cost amortization following the write-offs. Out of $35 million, $33 million could have been generated by a few strokes of the calculator.
While taking encouragement from the company’s cost-cutting efforts, Wall Street focused on the next promised bestseller. While we were not experts in this field, it appeared to us that analysts had already baked a very successful book into their estimates. Furthermore, we doubted whether the author, who was due to sign a new deal for the upcoming book, would settle for terms similar to those of the previous deal inked before she became such a runaway success. Management claimed otherwise, and asserted that the new deal would have similar
terms to the previous one. All told, we figured that most of the upside of the upcoming book was already incorporated into earnings estimates.
Finally, management was also promising organic revenue growth, even without the new bestseller. After looking at industry data, showing little to no growth over time, as well as spending a considerable amount of time trying to estimate the contribution of acquisitions, we came away feeling that the company was not really growing organically and could have a tough time doing so in the future.
We uncovered other disturbing trends. While earnings were strong over the previous few years, CFFO was lagging and free cash flow was basically nonexistent. Moreover, the company may have been over- reporting profits as capitalized prepublication costs had grown. These significant cash costs did not show up in the income statement, but rather found their way to the investing section of the cash flow state- ment below the CFFO line.11
We decided we had enough evidence to put on a small position. The following three quarters seemed to indicate that things were getting worse.
Other current assets began to build as did inventory, accounts receivable, and long term assets. The tax rate was declining, along with the inventory obsolescence reserve. Several lines of business showed signs of weakness, and the company began making contingent earn-out payments (to people who might still have been employed by the company) The stock price, however, continued to climb. Earnings were growing, and the hot author signed a new deal, although the terms were not disclosed. All was well as far as Wall Street was concerned. Throughout this period, we traded around the position but steadily increased it.
And then, almost a year after we began looking at the company, we got paid. The first sign of trouble was a slight EPS miss and lowered guidance that led to a nearly 10% drop in the stock during the day fol- lowing the announcement. The company blamed the miss on weak per- formance in a couple of product lines, and its lower guidance on delays in the new book release. The company did, however, guide for increased cost cutting to offset some of the EPS shortfall. The following quarter, the company preannounced, stating that it would report a loss in the upcoming quarter rather than the expected $0.31 profit, and guided down the next quarter as well. The stock tumbled 24% the following day. We covered our last share approximately 50% lower than where we initiated the position.
11By the way, we believe this is akin to how Tyco booked the costs incurred in sign- ing up new home security customers. Since, the company was buying the contracts from technically independent contractors it could book the associated installation costs under acquisitions.
Example 3: Failure—The Consumer Electronics Component Maker
The Consumer Electronics Component Maker showed up on our screens for cutting allowances for doubtful accounts, and building other current assets. In this case, the company actually reported pro forma EPS (aka EBBS) that were $0.07 better than expected in the most recent quarter.
The good numbers, combined with 10% of shares being short, led to a 20% move in the stock price following that release.
But the financial reports revealed that the “blow-out” quarter just reported might not have been so impressive. One could argue that lower allowances for doubtful accounts had boosted earnings by $0.04 per share. Other current assets had increased dramatically from the quarter before, hitting 9.5% of quarterly revenue from 5.5%. If the company were to keep other current assets as percent of sales flat on a sequential basis, EPS might have been $0.04 lower. In addition, the company sold some previously written down inventory that could have helped EPS by
$0.03. So the company might have had a total of $0.11 in what could be considered one-time EPS boosts.
Wall Street loved the company. Analysts and investors alike were infatuated with its strong unit growth. This was one of those “everyone will have one” stories as the company’s products found their way into many more consumer electronics goods. Already, they were selling into fast growing digital camera markets where there were reports of compo- nent shortages. Our research, however, uncovered significant supply additions planned within the next few quarters, as several competitors focused on gaining share. In addition, insider sales were significant and option costs were material.
We put on only a small position. We were still worried about the large short interest and also about our inability to accurately forecast the timing of supply additions in the face of very strong demand. Yes this was risky, but the rewards could have been huge. Wall Street tends to miss the inflection point when supply shortages (which force custom- ers to double order further exacerbating imbalances) turn into gluts (as increasing availability leads them to cut back on double ordering and draw down existing stocks) and the resulting EPS disappointments can be spectacular. We thought we might have had one here. The stock pro- ceeded to more than double over the next two months as the stock mar- ket ran. We slowly added to our position as we found more evidence that lots of supply was soon going to come online.
When the company reported the following quarter, it again blew by estimates as revenue soared. More concerning, to us, was the fact that the quarter seemed clean. We could not find any material indicators of EPS management in the financials. The stock, however, eased back to a level
only slightly higher than where we initially put on our position, indicating some rather lofty investor expectations had not been met. While we trimmed our position slightly, we still anticipated more downside.
When the annual report arrived on our desks (as that last reported quarter ended the fiscal year), we saw more EPS management signs, but we could not conclude how much they helped the latest quarter. We should have recognized that many of the symptoms that we saw previ- ously had disappeared and that it was time to realize a small loss. The combination of a high short interest and significant unit demand growth was a potent one. Yet we decided to stick around and wait for another quarter. That was a mistake.
The following quarter the company beat EPS by an even greater margin. Revenues were much better, and while we took issue with some of the line items, we had to agree that this was a strong quarter. The fol- lowing day the stock opened up over 20% and it never looked back.
We eventually admitted defeat after the stock had almost tripled from our entry. While prices for the company’s products were declining, unit growth was explosive and supply additions were coming on slowly. Our main mistake was waiting to see another quarter after the company reported a clean one in which it handily beat estimates. We paid a dear price.
SUMMARY
So much for our tales from the front. Sometimes you win and sometimes you lose. The key for us is to keep an even keel and not allow hubris to boost our levels of conviction and bet sizes during winning streaks; nor should we allow obstinance to keep us from realizing losses when we get some wrong. We could go on and on—we’re lots of fun at dinner parties—
but though the topic fascinates us, we doubt that we’d keep your attention.
In short, we try to pick our fights, and do so in a consistent manner in order that we might live to talk about them.12 We want to emphasize that this is only one strategy for short selling, and there are many viable and successful alternatives. Your approach should reflect both your ana- lytical and emotional strengths and weaknesses. The first three years of the new millennium gave too many people the impression that shorting was easy. This past year (2003) has shown that it can be rather hazard- ous to your financial and emotional well-being. Vigilance is key.
12Given our record on the short side in 2003, that “might” is still open to question.
Again, we remind the reader that since we run a somewhat balanced book, we may have a higher threshold for pain than some, as our longs might be benefiting from some of the same phenomena that have boosted our shorts.