October 27, 2011

Warranty Adjustments, Part 1:

When a company discovers it has more in its warranty reserve than is needed, it can withdraw the excess funds and add them to profits. Some of these financial adjustments can be huge, and can help dress up an otherwise disappointing financial report.

When claims are less than accruals, funds that could be better used elsewhere sit idle in the warranty reserve fund. When the excess is discovered, unneeded funds can be removed from the warranty reserve through what's called a change of estimate.

We took a look at the last ten quarters in the database, searching for instances where warranty providers withdrew funds from their reserves. These are instances where a company realized that they have over-accrued for warranty costs. They expected higher claims costs. But when those higher claims levels never materialized, they withdrew the excess funds from the warranty reserve fund.

Those funds are usually immediately converted into profits, and after taxes, they are added to net income. So this can get to be a very big deal, if the amount taken out of the warranty reserve is large enough. In one instance, in fact, the amount withdrawn turned out to be more than half that quarter's net income, which is another way of saying that a well-timed warranty adjustment doubled profits.

Changes of Estimate

In the ten quarters since the beginning of 2009, we found 914 favorable changes of estimate, ranging in size from a $153.00 adjustment made this year by Puradyn Filter Technologies Inc. to a $413 million adjustment made in 2009 by General Motors Co.

We tried our best to eliminate duplications, such as when a company makes a quarterly adjustment during the year and repeats the figures in its annual report. But some companies are finding their original claims cost predictions to be so pessimistic, or their reliability increases to be so phenomenal, that they're taking out funds year after year.

In Figure 1 below, we tallied the ten largest favorable adjustments since 2009, as measured in dollars. GM's $413 million change of estimate leads that list. That massive adjustment helped the company reduce the size of its warranty reserve fund from $8.491 billion at the beginning of 2009 to $7.062 billion by September 30, 2009.

Figure 1
Top U.S.-based Warranty Providers:
Largest Favorable Changes of Estimate
2009 to 2011
(Amount in US $ Millions)

  Company  Period   $ Amount   $ Reserves 
  General Motors Co. Sep09 -$413 $7,062
  Hewlett-Packard Co. 2009 -$223 $2,409
  Hewlett-Packard Co. 2010 -$53.0 $2,447
  Seagate Technology 2010 -$51.0 $372
  Motorola Mobility 2009 -$49.0 $156
  Motorola Solutions 2009 -$47.0 $209
  Garmin Ltd. 2010 -$42.8 $49.9
  United Technologies 2009 -$40.0 $1,072
  Hewlett-Packard Co. Apr11 -$38.0 $2,430
  Terex Corp. Dec09 -$35.2 $131

  Source: Warranty Week

Hewlett-Packard makes this list three times, because within each fiscal year it's been removing funds from its warranty reserve. In fiscal 2009, it removed $223 million. In fiscal 2010, it removed $53 million. And so far in fiscal 2011, it's removed $38 million. As of the end of April 2011, its warranty reserve balance was $2.43 billion.

That's a good thing, if it turns out that product quality is higher than was forecast, or repair costs are lower than was forecast. But it's not all good. It's a bit like predicting rain and getting a dry spell instead. Everyone's happy, but they're also carrying around umbrellas, and doubting your prescience. It would be better if accruals were always exactly enough to pay claims -- no more, no less.

The other repeat in Figure 1 is Motorola. Although the two separate companies Motorola Mobility Holdings Inc. and Motorola Solutions Inc. didn't formally exist until 2010, both restated their warranty accounts late last year as if they had been in business independently in 2009. And in that theoretical year apart, Motorola Mobility withdrew $49 million while Motorola Solutions withdrew $47 million from their reserve funds.

Sizeable Withdrawals

Seagate Technology plc, Terex Corp., and United Technologies Corp. were also among the companies making sizeable withdrawals from their warranty reserves since 2009. Understandably, they don't want to draw any attention to these transactions, for two reasons. First, they're correcting mistakes, where past estimates proved to be too high. And second, the suspicion is always there, that the company is somehow manipulating forecasts to provide magical boosts to earnings, just when they're needed.

We found little evidence of that. But we did look for ways to quantify these favorable changes of estimates in units other than dollars. For while GM and HP have hundreds of millions to move around, smaller companies don't. But proportionately, their changes of estimate can be even more beneficial.

In Figure 1, the final column lists the closing balance in the warranty reserve funds, after claims have been paid, after accruals have been made, and after these changes of estimate have been processed. For companies like GM and HP, the changes of estimate represented between 2/100ths and 9/100ths of their ending balance. In other words, even though the amounts taken out were large, the amounts left behind were even larger.

That's not so for all of them, however. Look again at Garmin Ltd. After removing $42.8 million from its reserves in 2010, the company had only $49.9 million left behind. More ominously, that amount was equal to what the company was paying in claims in only 4-1/2 months. But the company had also managed to reduce its claims rate significantly from where it was in 2008-2009.

Withdrawals Exceed Reserves?

Most shockingly, we found that a 1:1 ratio between the adjustment and the ending balance wasn't all that rare. In fact, of the 914 adjustments we found, 32 of them were greater than or equal to the ending balance. Garmin's adjustment was 41st on that list, the top ten of which are in Figure 2.

For instance, International Lottery & Totalizator Systems Inc., which as its name implies is a maker of lottery systems and gambling machinery, reported in the quarter ending in January 2009 that it withdrew $382,000 from its warranty reserve. That's no big deal until you notice that it left only $26,000 in its warranty reserve after that adjustment. And that means there is almost a 15:1 ratio between the size of the adjustment and the ending balance in the reserve.

Figure 2
Top U.S.-based Warranty Providers:
Largest Adjustments vs. Reserves
2009 to 2011
(Amount in US $ Millions)

  Company  Period   $ Amount   x Reserves 
  International Lottery Jan09 -$0.382 15
  Universal Electronics Jun10 -$0.063 7.9
  Altera Corp. Sep09 -$0.750 6.8
  AsiaInfo-Linkage 2010 -$0.253 5.8
  Key Tronic Corp. 2009 -$0.093 3.7
  Cabot Microelectronics Mar09 -$0.648 3.7
  Entropic Communications 2009 -$0.471 3.2
  Puradyn Filter 2009 -$0.109 2.7
  Monolithic Power Systems 2009 -$0.728 2.5
  Spansion Inc. 2010 -$3.21 1.8

  Source: Warranty Week

Notice that there are absolutely no overlaps between the top ten companies of Figure 1 and the top ten companies of Figure 2. Very large companies simply can't get away with removing half or more of their warranty reserves and adding that amount to profits. In fact, the only medium-sized companies we found that had taken an adjustment greater than or equal to the size of their warranty reserve since 2009 were Cymer Inc. and Centex Corp. Lam Research Corp. and Life Technologies Corp. came awfully close, but ultimately their adjustments were slightly smaller than their ending balances.

We also think another factor is at work here. Very large companies generally don't make very large mistakes -- at least not in the beneficial direction like we're detailing this week. What we're talking about here is a company suddenly realizing that half or more of their warranty reserves are unnecessary and superfluous.

Forecasting Accuracy

That's a wonderful discovery to make, but it's not one that a warranty management expert is going to be proud of making. It's like finding out you've been paying too much tax. If you're new, you're on your way to a promotion. But if you've been there for a while, it would be more beneficial in terms of job security if the adjustments were smaller and more frequent. And it would be best of all if the accrual rate turned out to be exactly right all along, making these adjustments unnecessary.

We should note that in 107 of the 914 adjustments, the company made no new accruals at the time they also made the adjustment. In some instances, the company listed a negative accrual, rather than a change of estimate and a $0 accrual. Both are examples of sloppy accounting methods which suggest a poor understanding of warranty management.

But let's get back to the positive. Let's assume that a company has been doing it right all along. It predicts a failure rate and a typical cost per repair. It determines the correct amount to accrue for each unit sold. And then it finds out that those estimates were a little high.

So while the company continues to set aside accruals with each new sale, the warranty manager determines that past over-accruals can be corrected by making a sizeable withdrawal from the warranty reserve. How sizeable? In Figure 3 were comparing the adjustments to the accruals. And for the top ten, the ratio ranges from 3.7:1 for Capstone Turbine Corp. to 35:1 for Terex Corp.

Figure 3
Top U.S.-based Warranty Providers:
Largest Adjustments vs. Accruals
2009 to 2011
(Amount in US $ Millions)

  Company  Period   $ Amount   x Accruals 
  Terex Corp. Dec09 -$35.2 35
  AsiaInfo-Linkage Mar09 -$0.129 26
  Puradyn Filter 2009 -$0.109 15
  Mattson Technology Mar09 -$0.729 12
  Aetrium Inc. Jun09 -$0.018 9.0
  M.D.C. Holdings Jun09 -$10.9 5.8
  Data I/O Corp. Mar09 -$0.025 5.0
  Nautilus Inc. Sep09 -$1.47 4.2
  Aehr Test Systems Aug10 -$0.088 4.2
  Capstone Turbine 2009 -$1.30 3.7

  Source: Warranty Week

In addition to the 107 instances where there were adjustments but no new accruals, there were also 110 instances where the adjustments were larger than the accruals. Terex has the highest ratio, but only because it dropped accruals to an artificially low level of $1 million during the final quarter of 2009. For the sake of comparison, claims topped $14.2 million during the same period.

Smaller Companies Dominate

And although two of the top ten are mid-sized companies -- homebuilder M.D.C. Holdings Inc. and exercise equipment maker Nautilus Inc. -- small companies again dominate the list. So there's Puradyn again. It's the same transaction, in a different comparison. The small auto supplier withdrew $108,533 from its warranty reserve fund in 2009, which doesn't sound like much until one notices that accruals were only $7,434 and claims were only $7,050 during the same period. Or to put it another way, Puradyn discovered an error equal to 15 years of warranty spending.

The irony is that a 15:1 ratio was good enough for only tenth place when we compared the size of the adjustments to the amount of claims. So don't cry for Puradyn's warranty manager, who had to tell the finance people that there was a 15x multiple between the annual claims bill and the over-accrual. Worry about the Alliant Techsystems Inc. warranty manager who had to explain a 91:1 error, or the AsiaInfo-Linkage Inc. warranty manager who had to explain a 168:1 error.

Figure 4
Top U.S.-based Warranty Providers:
Largest Adjustments vs. Claims
2009 to 2011
(Amount in US $ Millions)

  Company  Period   $ Amount   x Claims 
  AsiaInfo-Linkage Sep09 -$0.168 168
  Alliant Techsystems Dec10 -$1.27 91
  LaserCard Corp. Jun10 -$0.052 52
  Sycamore Networks Jan10 -$0.133 44
  Entropic Communications Mar09 -$0.185 31
  Finisar Corp. Oct10 -$2.58 30
  Foster Wheeler Mar09 -$11.3 23
  International Lottery Jan09 -$0.382 22
  Trio-Tech International Dec09 -$0.053 18
  Puradyn Filter 2009 -$0.109 15

  Source: Warranty Week

Incredibly, there were 22 instances where the ratio between adjustments and claims exceeded 10:1. And there were 173 others where the size of the adjustment equaled or exceeded the amount of claims. But there were also ten instances where the downward adjustments were accompanied by reports of $0 claims, therefore making a ratio impossible to calculate.

Can't Compare to Net Income?

We tried making a comparison between the size of the adjustments and the amounts reported as net income. But it failed, primarily because so many of these warranty providers declared net losses, especially during the recessionary year of 2009. Of the 40 companies on any of these top ten lists, only 20 were profitable in the period in which they made the adjustment, and two of those were profitable only because of the favorable outcomes of their bankruptcy reorganizations.

In only three instances were the adjustments more than 10% of the net income reported during the period. But in one of those instances, the adjustment managed to equal 58% of net income. And that would again be Garmin -- this time in the first quarter of 2010. During that quarter, it withdrew $21.8 million from its warranty reserve, and reported $37.3 million in net income. By the end of that year, as it continued to make withdrawals, the adjustments grew to $42.8 million but net income for the year grew even faster to $585 million.

At the other extreme, among the 20 companies declaring a net loss during the period in which they made a withdrawal from their warranty reserve, only International Lottery & Totalizator Systems made a withdrawal that exceeded the net loss. Motorola Solutions came close, and both M.D.C. Holdings and Presstek Inc. were on the high side. But they were still in the net loss column. In no cases did a warranty adjustment prevent a net loss from occurring.

Compared to Product Sales

Although we can't make a meaningful comparison to earnings, we can compare these adjustments to product sales. And we're happy to report that in all 914 instances where there were adjustments, there were also sales. Aetrium Inc., which made the top ten in Figure 3, was also among the smallest in terms of sales, with a $1.185 million total for the second quarter of 2009. H+GM, HP and General Electric Co. were the largest, in terms of sales.

In Figure 5, we listed the ten largest ratios between sales and warranty adjustments. But because these ratios were on the relatively small side, we used percentages. For instance, in the period in which International Lottery made a withdrawal ($382,000) that exceeded the net loss (-$199,000), product sales were above $4.7 million. To put it another way, the adjustment was equal to only 8.1% of sales, though it was almost twice as large as the net loss.

Figure 5
Top U.S.-based Warranty Providers:
Largest Adjustments vs. Sales
2009 to 2011
(Amount in US $ Millions)

  Company  Period   $ Amount   % of Sales 
  Terex Corp. Dec09 -$35.2 18%
  Mattson Technology Mar09 -$0.729 16%
  International Lottery Jan09 -$0.382 8.1%
  AsiaInfo-Linkage Mar09 -$0.168 6.0%
  M.D.C. Holdings Mar09 -$10.9 5.9%
  Palm Inc. Nov09 -$8.30 5.7%
  Puradyn Filter 2009 -$0.109 5.7%
  ZAP Jonway 2010 -$0.203 5.3%
  Presstek Inc. Mar09 -$0.258 5.2%
  Garmin Ltd. Mar10 -$21.8 5.1%

  Source: Warranty Week

The only new names on this chart are Palm Inc., Presstek Inc., and ZAP Jonway. Palm was a hand-held computer maker that's now part of HP. Presstek is a specialized computer printer maker. And ZAP Jonway (which acquired the Chinese automaker Zhejiang Jonway Automobile Co. Ltd. earlier this year) is an electric vehicle manufacturer.

Relatively Minor Adjustments?

Of the 914 adjustments we examined, there were only 134 instances where the adjustments exceeded 1% of sales. One can see from the results in Figure 5 that in only two instances did the warranty adjustments exceed 10% of product sales. And these are the only ten instances in which the adjustments exceeded 5% of sales.

In other words, while some of the downward adjustments were very large (Figure 1) or were large in comparison with warranty reserve balances (Figure 2), accrual totals (Figure 3), or claims totals (Figure 4), in only a small handful of cases did they even approach sales totals (Figure 5).

This doesn't mean there weren't companies that "made their numbers" by manipulating warranty balances. But it wasn't a widespread practice. And we couldn't find a single case where a favorable and well-timed adjustment meant the difference between profit and loss.

And while we shouldn't forget that every one of these adjustments is fixing a mistake, those corrections -- finding spare funds you didn't know you had -- occurred because warranty costs turned out to not be as high as expected. And whether that's because of bad math or good quality, it's still a positive development.

Next week, we'll examine the other end of the spectrum, where companies underestimated warranty expenses and had to make additional and sometimes painful contributions to their warranty reserve funds.

Assurant Solutions Shed The Light
Pegasystems Inc.
Fulcrum Analytics
GWSCA First Annual Conference on Service Contracts
Sign up for a free subscription to Warranty Week:
     subscribe     change of address     unsubscribe