After revenue recognition, inventory is the second most important factor for earnings quality analysis.
When a company creates product inventory, it invests cash. Management spends the Benjamins well or poorly, depending on its skill at judging business trends. The risk is that, the longer inventory ages, the greater the possibility that the company is misallocating cash, misjudging the market and may have to write down inventory. Therefore, inventory affects profitability via gross margin on the income statement and shows demand for a company’s products.
We analyze inventory numerous ways on the balance sheet to get a better sense of the impact it may have on the company’s product demand. Similarly to DSO, we track inventory on a days sales basis and look for both year-over-year and sequential changes:
Let’s turn to John’s notification to clients about semiconductor maker Silicon Laboratories. In the quarter in which he flagged the issue, days sales in inventory (DSI) jumped dramatically.
Silicon Laboratories experienced a spike in its inventory levels in the June 2004 quarter when inventory grew 155 percent year-over-year compared with sales growth of 83 percent. Table 3.1 shows the result, that days sales in inventory (DSI) jumped 14 days year-over-year to 55 days. On a sequential basis, DSI increased five days.
Table 3.1 Silicon Laboratories’ Increasing DSI
During the June 2004 quarter conference call, Russ Brennan, Silicon’s CFO, put the inventory situation in a positive light. According to Brennan, inventory at the finished-goods levels was depleted and “inventory distribution also went up consistent with our plan to increase MCU [microcontroller unit] sales and distribution infrastructure to support this ramping business and to support projected growth in our other broad-based, mixed-signal businesses.” Despite this explanation, the next speaker, CEO Dan Artusi provided an outlook for the third quarter with revenue in the range of $120–$123.6 million and EPS of $0.36–$0.38, below Street expectations of $129 million and $0.40, respectively. Economic issues related to China during the quarter were in part to blame, but we think China is a longer-term probem that extends beyond one quarter.
The inventory analysis extends beyond the amount reported on the balance sheet. In general, management is quick to point out conservative revenue recognition policies, whereby revenue is not recognized until goods are sold through to the end user. In addition to inventory on the balance sheet, the company also records deferred income to distributors, to whom it affords rights of return and price protection for products they don’t sell. This revenue recognition is deferred until the product is sold to the distributor’s end customer. While deferred income represents some future revenue, in Silicon’s case it also represents inventory sitting in the sales channel that has yet to be sold and is at risk. Table 3.2 illustrates the change in deferred income and in adjusted DSI at Silicon.
Table 3.2 Silicon Laboratories’ Deferred Income to Distributors Worsens DSI*
Viewed this way—with deferred income showing potential inventory at risk—the inventory situation is worse. While this revenue may be recognized if and when the distributor sells through to the end customer, DSI reflecting this inventory still in the channel jumped seven days sequentially and nineteen days year-over-year. Distributors may slow purchases from Silicon Laboratories until demand picks up and the goods are sold through, or the company may have to cut prices to move the goods through the channel, slicing profit margins.
Management admitted to excess supply in the sales channel with respect to demand for mobile handsets that use the company’s chips. If the situation, in China in particular, does not turn around quickly and results in a protracted slowdown for the mobile market—which we think is a possibility—then the inventory situation will worsen before it gets better.
To be sure, investors are aware that inventory spiked in the recent quarter. The question is whether this is a one-quarter blip or a major issue. Negatively, technology research firms have consistently overestimated cell-phone demand…. Given that the market researchers have consistently been wrong with respect to consumer demand, there is a risk that the inventory build-up situation across all of these companies—including Silicon—is a significant long-term risk.
We can dig deeper for information on whether inventory issues are a blip or persistent. Inventory is not one undifferentiated blob. For example, with a soft-drink company, it can start with the sugar or high fructose corn syrup (horrors) and other raw materials, then treating these with the secret-formula mixing process (work in process) and then canning or bottling the concoction as finished goods ready for distribution. It doesn’t matter that, in today’s world, parts of this process are performed in dozens of facilities at different times throughout the world. Management still has to control the process in a manner beneficial and transparent to shareholders. Not all companies do. So, while it’s important to look at inventory as a whole, companies show their management expertise and earnings quality in all stages of the inventory production process, sometimes revealing important trends to the alert investor.
We now mine the raw inventory number on the balance sheet because it may be insufficient to detect earnings quality issues for companies that sell physical products requiring inventory production stages (even the often-used example of the lemonade stand has these definable stages). Companies typically, though not always, report raw materials, work in process, and finished goods. The best list them under inventory on the balance sheet, but for the rest we find the information in the notes to the financials. And some, unfortunately, do not break out the components at all.
We then can calculate the DSI for each component and compare it to sales growth:
In Table 3.3, we use component information found in the footnotes for semiconductor vendor Intel to calculate DSI for each component and find excellent inventory management.
Table 3.3 Intel’s Efficient Management of Work-in-Process Inventory Drives DSI Improvement: December Quarter 2009–October Quarter 2011* DSI for Inventory Components
From Q4 2009 through Q2 2010, Intel’s DSI rose well in excess of revenue, suggesting poor inventory management. The next two quarters’ difference was unremarkable, but from there the company ran like an atomic clock. Revenues rose while inventories shrank. The efficiency came almost entirely from managing work in process from DSI of 43 to 26. Raw materials and finished goods remained the same, despite rising sales, but work in process moved faster.
Tracking the components here shows nothing but good at nimble Intel. During these two years, it was briefly caught off guard before it righted the chip—in an uncertain economic environment to boot. So be alert. More dramatic divergence from a company’s normal inventory component pattern may be good or bad. It alerts the investor to look more closely to determine whether there are earnings quality issues and to see both good things or bad things coming down the road before others do. Next we’ll explain that in more detail.
As with Intel, a well-run company’s inventory components are consistent with a company’s need to manage the ups and downs of revenue. They can diverge from a seasonal pattern or usual business conditions for either positive or negative reasons. A company receiving new orders knows to build inventory components to fulfill them, and the first positive sign may be a rise in raw materials followed by work-in-process: they diverge positively . Conversely, inventory components may bulge through poor management and poor demand, diverging negatively .
Consider the snake. It’s good digestive inventory management when it snares its rodent dinner. The bulge nearer the head—raw material—promises good things to come along the digestive track through work-in-process and finished goods. Yum. It’s efficient and only eats what it needs, and when it does, it’s positive food inventory divergence.
Now consider a person at Thanksgiving, eating a meal many multiples more calories and sheer volume than usual. It takes far, far longer to digest (aging inventory), and, turkey’s sleep-inducing tryptophan aside, there is the risk of inability to digest many nutrients and perhaps much of the meal at all (wasted inventory—call it a buildup of finished goods). Heartburn ensues.
Tracking inventory component divergence gives the investor advance notice of good or bad performance. Yet, few investors monitor these components, even though they can be the difference between investment profit and pain. We owe this subtle and very profitable concept to quality-of-earnings pioneer Thornton (Ted) O’glove, who presents positive and negative inventory component divergence in his path-breaking Quality of Earnings .1 We will leave out “component” to call this simply positive or negative divergence.
At its root, positive divergence occurs when a company knows that business ahead is good, so its raw material inventory picks up while work in process and finished goods either lag or do not increase at the same rate. This can be useful to identify an investment opportunity in many places, but especially in a cyclical business or a turnaround situation. With positive divergence, companies in these situations likely see better cash flow, earnings, and stock price ahead.
The investor wants to avoid the opposite: when finished goods pile up in excess of sales while raw material inventories and work in process are flat or declining absolutely or relative to sales. This negative divergence is a good predictor of potential inventory write-downs and poor earnings, cash flow, and stock price.
Inventory component divergence is less amenable to analysis than many other indicators because some companies provide components annually, not quarterly, in the footnotes and not on the balance sheet, or not at all. To focus initial research, O’glove in 1987 emphasized that the investor should concentrate on “industries subject to rapid changes in products and taste” such as “high fashion, seasonal goods, and high tech.”2 This is sound advice, though the investor with expertise in other industries can do inventory component analysis there. The bottom line: Focus on negative divergences wherever you find them and understand the business.
While high fashion can certainly be a fertile hunting ground, any retailer subject to fashion tastes and trends can experience ballooning inventory through the slightest error in estimating demand. Even the best management can have trouble predicting fickle public taste for established, let alone new, apparel.
A useful example of misjudging retail demand is shoe company Crocs, whose stock plummeted in late 2007 into 2008, ahead of the broad market crash. DSI for finished goods and in relation to sales gave plenty of warning. Table 3.4 shows that raw materials and work in process were flat while finished goods soared.
Table 3.4 Crocs’ Negative Inventory Divergence: June Quarter 2007–September Quarter 2008*
Dramatic negative divergence of finished goods to sales sequentially is evident from the third quarter of 2007 to the first quarter of 2008. DSI exploded versus revenues from the second quarter to third and fourth of 2007. DSI still rose in the first quarter of 2008, and then the company slammed on the inventory brakes, with DSI dropping by 67 percent in two quarters. People were—and are—buying the shoes, but the company got demand wrong, overinvested in inventory, and brought investors great pain. Warehouses were chock-a-block with crocks of Crocs.
This is a reminder that avoiding a blowup is one thing, but to profit from it does not require perfect timing. Short sellers often mistakenly anchor to the highest recent price and think they “missed it” (just as longs can anchor to the recent lower prices and rue their timing). Remember that on the short side you make as much on the downside if a stock goes from 100 to 50 as from 50 to 25, 10 to 5—you get the idea. All those still provide 50 percent gains. For Crocs, it’s the third quarter of 2007 when exploding DSI appeared in the 10-Q filed November 14, 2007, and the fourth quarter DSI jump in the 10-K filed February 29, 2008. There was money to be made, not only shorting on the first DSI in the 10-Q, but also acting on the second, the 10-K information over three months later. (See Figure 1.3 in Chapter 1.)
Crocs offers an unusually clear case where inventory component analysis could have led an investor to avoid, sell, or short Crocs at appropriate times. There are equally important cases where the analysis is subtler.
For so-called “tech” companies that make physical products, the longer their inventory collects dust, the sooner the company or competitors incorporate advances so that aging inventory either becomes obsolete and unsellable at any price or brings declining average selling prices that quickly eliminate profit margins. This competition risk is so intense that some investors avoid altogether so-called “physical tech” product companies, where price deflation is the norm and volume has to be growing for revenues and profits to stay in place. (They heed Berkshire Hathaway Vice Chairman Charlie Munger’s advice to investors that technological advances bring products and favorable declining prices that benefit consumers and not shareholders of the producers.)
John’s June 27, 2005 report on analog circuit maker Maxim Integrated Products shows negative inventory divergence predicting poor earnings quality and trouble ahead. Maxim buys silicon-wafer raw materials or manufactures wafers itself. Work in process consists of manufacturing the chips on the wafers (also called die) and determining how many of them form the “die bank” inventory—the good die that can be parts of more product or shipped to customers for their use.4 John advised clients that raw materials DSI did not keep pace with the rise in DSI for work in process and finished goods.
Inventories on the Rise
We are concerned with Maxim’s recent high level of manufacturing output, which has not only kept gross margin elevated, but has increased inventory levels substantially. In fact, inventory levels, mostly in die-bank inventory, have increased significantly and are expected to rise in the near future. In the March 2005 quarter, inventory, as measured by days sales in inventory (DSI) increased to its highest level in the past eight quarters. As illustrated in Table 3.5, DSI increased 18 days sequentially and 46 days year-over-year to 130 days in March .
Table 3.5 Maxim Integrated Products Negative Inventory Divergence, from Fourth Quarter of 2003 through Third Quarter of 2005*
We also note the negative inventory divergence in the March 2005 quarter with an increase of 59 percent, and 31 percent year-over-year, in both finished goods and die bank respectively, compared to an increase of only 12 percent in raw materials. In our opinion, manufacturing continues well beyond necessary to satisfy future demand. Should demand not grow as expected, we can expect a future decline in margins.
Table 3.5 shows that finished-goods DSI increased 6 days sequentially and 18 days year-over-year, compared to flat sequentially and a 4-day year-over-year increase for raw materials DSI. However, the company attributes the majority of the DSI increase to higher die-bank or work-in-process inventory, which has risen 12 days sequentially and 23 days year-over-year.
According to Maxim, the buildup in die bank has been necessary to manage lower lead times as well as anticipated increased demand. But as Table 3.6 reveals, Linear Technology, Maxim’s most direct competitor, maintains DSI of only 53 compared to Maxim’s 130 days with equivalent lead times of 6 weeks, strongly suggesting Maxim’s die bank expansion may be unnecessary. In fact, Maxim’s lead times have contracted significantly from 12 weeks six months ago to 6 weeks today, but that improvement isn’t expected to last. Management believes “lead times of six weeks [are] likely a bottom.” Should lead times return to normal levels, the excess die-bank inventory would be unjustified at best.
Table 3.6 Maxim Integrated Products’ Worsening DSI, from Fourth Quarter 2003–Third Quarter 2005*
Because they think that die-bank inventory does not generally become obsolete, Street analysts aren’t concerned about this particular inventory growth. Lost on them is that there is quite a distance between obsolete (zero value) and full value, and the company has a tendency to write off inventory, dragging down margin levels. In fact, inventory write-downs of $5.0 million and $11.9 million in 2004 and 2003 respectively reduced gross margins both years, and those write-downs included both work-in-process and finished-goods inventory. This indicates that excess die-bank inventory poses some risk of obsolescence.
We also believe that these high inventories threaten continued production at current rates. Given that depreciation is nearly half of the cost of goods sold for Maxim, spreading that high fixed cost over lower output would dramatically reduce gross margin. As a result, we think the high gross margins and inventory levels augur Maxim’s future earnings disappointment.
Figure 3.1 shows the story’s sad end. When John issued this report on June 25, 2005, Maxim shares stood at $38.83. In six months, they dropped 20 percent to around $30 while the broad market advanced. Avoiding that loss was good because shares haven’t traded above $30 since, but there’s more. Paying attention to these warning signals also avoided a devastating loss when, in late 2007, the company reported a restatement and went onto the pink sheets on October 2 at $28.26. When the stock returned to the Nasdaq a year later, the first-day close was $13.73. Investors aware of the issues avoided or profited from losses.
Figure 3.1 Maxim Integrated Products: 2004–December 9, 2011.
Source: FreeStockCharts.com.
Analyzing inventory components offers positive and negative information. Overall, though, rising inventory by itself is no indicator of problems. What matters is inventory buildup relative to other factors.
Investors should watch closely the rate of buildup versus sales. Inventories may rise for good or bad reasons, but rarely do analysts consider that lower inventory buildup—shown also in decreasing accounts payable—may mean that existing inventory ages while the lower percentage of new, presumably more market-responsive inventory is most likely to sell. If aging inventory is a greater percentage of total inventory, then mark-downs, write-downs, and obsolescence have a more severe impact. This is especially troubling with consumer products—companies from whom customers demand today’s products in new packaging.
Inventory buildup also can be a problem for newly acquired companies. When one public company acquires another public company, examining the target’s inventory buildup can show trouble ahead for the acquirer. In November 2006, John issued a report to clients that, among other things, noted VeriFone’s risk from substantial inventory buildup at its acquisition of Lipman Electronic Engineering. As Table 3.7 shows, in each of the four quarters prior to the report, inventories surged on a year-over-year basis, and DSI grew 81 percent from 90 days in the June 2005 quarter to 163 days in the June 2006 quarter.
Table 3.7 Inventory Buildup at VeriFone’s Lipman Acquisition: September Quarter 2004–June Quarter 2006*
While inventories ballooned, Lipman’s gross profit margin remained relatively stable. This masked underlying weakness in Lipman’s profitability, because the margin benefited from increased production spreading fixed costs across more product units. The company stated in the footnotes to the financial statements that it experienced a loss of pricing power and that it built up inventory in expectation of sales that—particularly in one subsidiary—never materialized. However, despite the disappointing sales, the company had not written off the inventory. That risk to future quarterly margins remained.
VeriFone shares closed at $32.15 on November 13, 2006. Shorting took some patience: The shares rose during the next year, but the company could avoid the inevitable only so long. On December 3, 2007, the company announced that it would restate quarterly financial statements for the nine months before July 31, 2007. The stock price dropped 46 percent that day and extended the loss to 58 percent in three days.3 Seeing the inventory buildup gave alert investors ample time to sell the stock or short for profit.
In a system where revenue is recorded when a sale is made, the company must figure out the cost of goods sold by taking a physical inventory. There are three methods: LIFO (last-in first-out), FIFO (first-in first-out), and WAC (weighted average unit cost). FIFO bases costs chronologically, LIFO reverse chronologically, and WAC on the weighted averaged for the period. An investor wants consistent treatment, so that comparisons are meaningful. Changes from one method to another for any purpose, such as valuing reserves, are cause for concern.
The LIFO reserve is the difference between the FIFO and LIFO cost of inventory for accounting purposes. The LIFO reserve is an account used to bridge the gap between FIFO and LIFO costs when a company is using FIFO but would like to report LIFO in its financial statement, which it is allowed to do. When input prices decline (“first in”), the reserve is reduced and the cost of goods sold (COGS) declines. But management can also dip into the reserve by reducing LIFO inventory and use it as a source of earnings. When a reserve is reversed, it’s a credit to the income statement. Essentially, it’s 100 percent margin, so the EPS boost is significant. A reduction in the reserve for no apparent reason is a red flag.
If a company reverses or lowers its LIFO reserve, it may cause an unusual change in gross profit margin. Without a plausible explanation, management may be boosting EPS stealthily.
While LIFO accounting is probably the best understood of the inventory accounting concerns, consider the situation of impaired inventory. Generally accepted accounting principles (GAAP) require that inventory be written down to fair market value when the impairment is realized. If the company later sells that reduced-value inventory, it can see a jump in gross margin, because much of the cost of goods sold related to that merchandise was booked in the inventory charge. So Wall Street analysts add back a one-time impairment charge, but they do not note that future profits are generated as a direct result. When they “forget” to adjust future reports accordingly, results look better, but they’re one-time only.
Here’s another about-face on inventory: writing it off today only to sell it down the road.
It seems impossible, but a company can write off inventory now, yet still sell it later. This is among the most aggressive inventory actions management can take. When a company writes off inventory in one quarter as obsolete and then sells it in another quarter, it artificially overstates margins. As with all write-offs, if a company has more than one it may well “kitchen sink” them, taking a big hit all at once rather than dying the death of 1,000 cuts. If so, the overstated benefits when then-written-down inventory is subsequently sold appear even more dramatically but unsustainably “good.”
This happened frequently during the 2000–2002 crash, when many companies that made physical technology products built up inventory, expecting skyrocketing demand to continue. Sometimes, they wrote it off and sold it later. In 2001, for example, networking equipment maker Cisco Systems recorded an excess inventory charge of $2.25 billion (a write-off) based on future sales forecasts, caught by the recession with declining demand in an industry where technology changes rapidly.
In the dramatic explosion of telecommunications networking equipment for Internet network buildout, Cisco’s inventory zoomed from $655 million at October 30, 1999, to $2 billion a year later, reaching an all-time high of $2.5 billion a quarter after that. No wonder Cisco took a startling $2.25 billion write-down all at once to take a one-time hit that allowed it to “get it all behind.” The company stated: “This additional excess inventory charge was calculated based on the inventory levels in excess of 12-month demand for each specific product. We do not currently anticipate that the excess inventory subject to this provision will be used at a later date, based on our current 12-month demand forecast.” Yowza!
But it didn’t take long to partially revise that forecast. In subsequent quarters, Cisco used and sold some of the inventory, reversing the charge and benefitting the income statement, as Table 3.8 shows.
Table 3.8 Cisco Systems’ Sale of Written-Off Inventory: 2001–2002*
Of the $2.25 billion reserve recorded in FY 2001, 32 percent—$712 million—was in effect later sold. The company actually sold $23 million to outside customers, used inventory internally for a gain of $457 million in operating cash flow, and settled for $252 million with customers who probably cancelled orders.
The reversal of reserves through sale, utilization, and settlement appears to boost revenues, but only because of the prior severe write-down. This creates the appearance of improvement where there would have been none. Moreover, the magnitude of the “inventory utilized” and the uncertainty over valuation for internal use makes it hard to have any confidence in these numbers, operating cash flow, margins, or EPS. The best advice is to avoid—not even short—such a stock.
Here, a company may change from average cost to retail method to value inventory. The retail method recognizes markdowns as the products are sold, while the average cost measures markdowns when made.
Consider mythical Teenybopper Inc. that sells tops. It realizes it’s ordered far more of its St. Swithin’s day tops than it can sell in that short holiday season. It puts a clearance sign on the display for the tops announcing a markdown from $15 to $10. The conservative and preferred method is average cost, which recognizes the markdown, therefore reducing inventory valuation, when made—when the sign is put on the display and tops, if sold, would be rung up at the register at $10. The aggressive practice is the retail method, where the store puts up the sign, but only recognizes the markdown and reduces inventory value retrospectively, when the top is rung up at the register at the marked-down price.
The change from average cost to retail method stretches out the period between markdown and actual reduction in inventory value. The average cost method is more accurate. The inventory is worth less the moment it’s marked down; the top isn’t worth $15 anymore. In fact, it may actually be worth even less than $10, because in retail, the first markdown is rarely the last, until the last stop for the hapless St. Swithin’s day tops is the discount outlet mall (Tom’s favorite shopping destination).
An inventory step-up is when a company makes an acquisition and increases—“steps up”—the value of the acquired company’s inventory on the balance sheet. When the inventory is sold, the stepped-up cost is expensed as a special, nonrecurring item. A March 2008 Assay Research report showed these problems at Jarden, a diversified consumer products company:
We question the quality of Jarden’s reported gross margin improvement over the past two years, given that the Company excluded manufacturer’s profit in inventory—the purchase accounting fair value adjustment to acquired inventory—from the calculation. When a company makes an acquisition, it records the acquired assets, including inventory, at fair market value, which can be above book value. In those cases where inventory is “stepped-up,” future cost of sales would be higher, since the carrying value of the acquired inventory was adjusted upwards at the time of the deal when the acquired inventory is sold. During nine out of the past twelve periods, Jarden has adjusted its gross margin upwards by excluding the portion of reported cost of sales attributable to this inventory step-up [emphasis added]. Jarden reported that its adjusted gross margin increased year-over-year by 220 basis points in the December and September quarters, and by 350 and 100 basis points during the June and March periods, respectively [see Table 3.9]. Because of the frequency and recent significance of the charges [see Table 3.10], we are concerned that Jarden’s gross margin may not have improved as much as reported. 4
Table 3.9 Jarden’s Gross Margin “Improvement”: March Quarter 2006–December Quarter 2007
Table 3.10 Jarden’s Quarterly Adjustments to Costs of Goods Sold for Manufacturer’s Profit in Inventory, 2005–2007*
These adjustments suggested that Jarden’s gross margin improvement in 2007 was unsustainable. While 2008 was harsh to all stocks, Jarden nevertheless dropped 60 percent from March 1, 2008, to its March 9, 2009 bottom, 12 percentage points more than the S&P 500’s fall.
Inventory management is not easy, so when a company does it well—knows its customers, seasonality, and product risks—it shows high earnings quality. When management gets it wrong, inventory analysis can show trouble ahead. But when they’re not just “getting it wrong”—when a company hides the deteriorating situation and poor earnings quality—that’s the ticking bomb this chapter’s tools will help you avoid.
Aggressive revenue recognition and inventory management are the top two predictors of poor earnings quality, but others compete for the title. We move now to other ways companies unsustainably boost earnings.
