Top 50 Warranty Providers:

U.S.-based manufacturers seem to have gained some control over their warranty expenditures during the past two years. Of the top 50 warranty issuers in terms of dollars spent satisfying claims, 12 were able to spend less in 2004 than they did in 2003, and another 22 spent a lower percentage of their sales revenue on warranty claims. That means only one out of three spent more dollars and/or a greater percentage of sales on warranty in 2004 than they did in 2003.

Overall, the warranty claims total for 765 different manufacturers tracked by Warranty Week in calendar 2004 was $25.1 billion, up 4.8% from calendar 2003. Warranty accruals grew an astonishing 11% to $27.2 billion, causing the combined total for the 765 separate warranty reserves to approach $37 billion.

Because of all the coverage of the Warranty Chain Management conference in recent weeks, the normal schedule for the annual roundup of trends in product warranty expenditures is almost a month late. Typically, most manufacturers have filed their annual reports by the end of March, allowing Warranty Week to publish the results in early April. Normally, there are a few stragglers who don't file their Form 10-K reports with the U.S. Securities and Exchange Commission until mid-April, which means we have to make a few estimates. However, because this year's edition is late, all but only two of the top 50 warranty providers have filed their reports, and both of those are late because of internal accounting chaos and not because of the calendar.

American manufacturers are now spending more than $3 million an hour on warranty claims. Claims rose from $6.225 billion in the fourth quarter of 2003 to $6.608 billion in 2004. That's more than $73 million per day. However, because product sales grew at more or less the same rate as product warranty claims, thanks to continued economic expansion, the overall claims rate was more or less unchanged. It stood at 1.84% of sales at the end of 2003 and at 1.80% at the end of 2004.

The Big Four

In 2004, four companies reported more than $1 billion in warranty claims: GM, Ford, HP, and Dell. Their claims rates, which are computed after making all the appropriate subtractions for sales of services, finance, software, and consumables, ranged from 2.5% for Ford to 3.7% for HP. The claims rate for Dell, however, was based upon only an estimate for sales of warranted computer hardware, since the company does not segment its revenue by type.

Another 27 companies reported between $100 million and $1 billion in claims; and 156 companies reported between $10 million and $100 million in claims. Four companies reported $1 billion or more in warranty accruals; 29 companies reported between $100 million and $1 billion in accruals; and 160 companies reported between $10 million and $100 million in accruals. The vast majority of manufacturers reported less than $10 million per year in either claims, accruals, or both.

The claims rate is computed by dividing the dollar figure for warranty claims by warranted product revenue. The accrual rate is computed by dividing warranty accruals by warranted product revenue. While the overall weighted average for the claims rate is around 1.8%, the median is closer to one percent. That means roughly half of all manufacturers have claims rates above 1%, and half are below 1%. However, many of the largest are well above 1%.

This is the case because relatively few large and well-known customer-facing brands seem to absorb the bulk of all warranty expenses, while the bulk of their suppliers and competitors seem to get off rather lightly. In other words, a well-known Detroit-based passenger car company whose name goes on the back of the vehicle can expect a claims rate around 2.5% to 3.5%, but its suppliers will typically pay only 0.5% to 1% of their revenue.

Wide Variance in Rates

There is, however, a very wide variance among all manufacturers. Claims rates for the top 50 warranty providers ranged from a low of 0.4% for Tyco to a high of 8.2% for Lexmark. Accrual rates ranged from 0.1% for Tyco to 11% for Lexmark. One reason Tyco's rates were so low was because the company does not divide its reported $7.79 billion in product sales into segments for warranted and non-warranted products. Therefore, we have probably included more product revenue than we should have in the denominator of the percentage calculations.

One reason Lexmark's rates were so high was because fully 63% of its revenue came from the sale of printer ink, toner cartridges, ribbons, and other consumables. Although they're replaceable if defective, consumables are usually not covered by express warranties. It's the same with light bulbs, brake pads, windshield wipers, and vacuum bags. When those products are working normally, they wear out and are replaced. That's not a defect. Warranties cover defects.

That having been said, very few companies besides Lexmark report their revenue in a way that allows for easy subtraction of non-warranted products from the total. Several of the top 50 provide little or no revenue segmentation data at all. In those cases -- specifically Ingersoll-Rand; United Technologies; Gateway Inc.; York International Corp.; American Standard Companies; Whirlpool Corp.; Brunswick Corp.; Nortel Networks Corp.; Maytag Corp.; and Dell -- estimates were made for the percentage of total revenue attributable to sales of non-warranted products, services, software, insurance, etc. This had the effect of raising the claims and accrual rates above what they would be if total revenue were used in the denominator.

Constant Methodology

In all cases, the methodology for each company's estimate was kept the same from one year to the next, so fluctuations in claims and accrual rates were due entirely to other factors. For instance, at Gateway 10% of total 2004 revenue was estimated to come from the sale of services, finance, insurance, etc. This percentage was the same in 2003. Therefore, when claims fell from $154 million in 2003 to $63 million in 2004, and when product revenue (after subtractions) rose from $3.06 to $3.29 billion, the claims rate fell from 5.0% to 1.9%.

If a company radically changes its warranty accounting methodology, it is supposed to disclose this information. So, if one year a company decides to include the cost of parts but exclude labor costs, and the next year it includes labor but excludes parts, it is supposed to let people know about this change. Because of the absence of these disclosures, we're going to assume constant methodology for the manufacturers as well, even though some reported incredible rate changes.

By the way, all the data was computed for calendar 2004, disregarding each company's specific fiscal year. For companies whose fiscal year ended on a date besides Dec. 31, the quarter that ended any time between Oct. 1 and Dec. 31 was counted as their fourth. In other words, Sun Microsystems ends its fiscal year on June 30, so its second fiscal quarter coincides precisely with the fourth calendar quarter. Dell ends its fiscal year on Jan. 28, so its third fiscal quarter ended Oct. 29 was used as the fourth calendar quarter in our calculations. The only exceptions to this rule were those few companies whose fiscal years ended on Jan. 1 or Jan. 3. Those were close enough to Dec. 31 to be left as is.

The Biggest Balances

Warranty reserve balances are normally reported by each manufacturer for the beginning and the end of each reporting period. To the beginning balance, accruals are added and claims are subtracted. After a few other adjustments for acquisitions, foreign currency fluctuations, and changes in estimate, the ending balance is calculated. Normally, manufacturers present this data as a table in the footnotes of their financial statements.

At year's end, six companies held $1 billion or more in their warranty reserve: GM, Ford, HP, Dell, GE, and United Technologies. Another 36 companies held between $100 million and $1 billion in reserve to pay future warranty claims; and 166 companies held between $10 million and $100 million. The rest held $10 million or less in their warranty reserves.

The dollar amounts are huge, but what's most relevant to each company is the percentage of sales that these dollar amounts represent. A claims or accrual rate above 5% is usually taken as a sign of trouble, except if the business model emphasizes the sale of more profitable services or supplies, as is the case with printers. But because different companies count up the components of warranty in different ways, one manufacturer's 5% may not have the same meaning as another's 5%, especially if they're in different industries.

It is therefore unwise to try to compare claims rates directly. What would be more revealing would be to compare a company's claims rate a year ago to its claims rate at the end of 2004. Did it go up? Did it go down? At a deeper level, did it go down primarily because of an increase in sales, or because of a decrease in dollars spent on claims?

The Top 50

Of the top 50 warranty providers, 34 saw claims as a percentage of product sales decline in 2004, while 16 saw their claims rate rise. That is more or less in line with the changes observed for all manufacturers. Overall, 63% saw their claims rate fall, 35% saw it rise, and 2% had it remain unchanged from 2003 to 2004.

At a deeper level, 22 of the 34 who saw their claims rate fall actually paid out more in claims in 2004 than they did in 2003. But their sales volumes rose faster than claims, so the overall percentage fell. All 16 of the companies that saw their claims rate rise also paid out more in 2004 than in 2003. Only 12 of the 50 companies on the list below paid out less in 2004 in terms of both dollars and percentage of sales.

Note: (1) -- Terex has been unable to file an
annual report; estimate based on six-month data.

The list of non-reporting companies was set to include Nortel Networks, but the telecom equipment maker finally filed its 2004 annual report with the Securities and Exchange Commission on Monday, May 2. Delphi finally came through with its remaining 2004 financial reports at the end of June 2005. Therefore, the only non-reporting company on the list above is Terex. We suspect there are other manufacturers who are issuing product warranties but who are not making the required disclosures about them in their financial statements. But we'll deal with them in a separate article.

Compared to Themselves

The chart above compares 50 manufacturers against themselves. It shows where they are now, where they were a year ago, and which direction they're heading in. Those with the largest percentage decreases are at the top of the list and those with the largest percentage increases are at the bottom. Admittedly, this doesn't reveal much, because a company that went from 4.1% to 4.0% would be ranked the same as one that went from 0.2% to 0.1%.

Here's another idea: Let's assume that the point at which each company began 2003 is equal to 1.0. If their warranty claims rate were to suddenly surge 25%, that value would increase to 1.25. If it were to suddenly plummet by 25%, that value would fall to 0.75. Last week, we included a chart that showed the changes calculated for the Big Three automakers of Detroit. This week, let's make the same chart for the top 50 warranty providers.

In the chart below, each of the 50 companies is represented by a line. For each line, the claims rate as of the beginning of 2003 equals 1.0. If the claims rate subsequently doubled -- for instance, if it went from 2% to 4% -- the line would move from 1.0 to 2.0.

In at least one case, the line actually passed 2.5. This means that the claims rate grew by a factor of 2.5. In this case, the company is Nortel. In early 2003 its claims rate was 1.67%. By late 2004 it had gone over 4.44%. So yes, it has grown by a factor of greater than 2.5. Since our scale stops at 2.0, the line for Nortel flies off the chart.

Compared to Each Other

We're going to color the line for Nortel red, along with the lines for Lennar, Motorola, and Honeywell. It is clear that these four companies have seen their claims rates rise the most since the beginning of 2003. Likewise, we're going to color the lines for Lucent, Gateway, Applied Materials, and Cummins green. They have seen their claims rates fall the furthest since the beginning of 2003. The lines for these eight companies look like the frayed threads of a rope -- clearly they're the exceptions.

Top 50 U.S.-based Warranty Providers
Fluctuations in Claims Rates, 2003-2004
(Jan. 1, 2003 = 1.0)

The other 42 companies are somewhat lost in the tangle of the black rope in the middle. They have seen their claims rates change by plus or minus 40% or less. That's actually quite a spread for only two years, but as the chart shows, the four companies in red saw claims rise by 40% or more while the four companies in green saw claims fall by 40% or more.

If we changed the scale to just plus or minus 40%, that is, 0.6 to 1.4, more companies that saw relatively large changes in their warranty claims rates during the past 24 months would be visible, but those with the biggest changes of all would fly off the edges. Successful warranty cost reductions by Deere, Whirlpool, and Tyco are lost in the tangle, as are slight increases for Apple, Kodak, and Delphi. The point of putting so many lines on top of each other is to show that there really is a "normal" range that holds most companies: plus or minus 30% or so. That makes the anomalies all the more obvious. They look like loose threads on the rope.

The scale also could be changed so that the end of 2003 equals 1.0, which would array the companies into more of an X pattern. Or it could be changed so that the end of 2004 equals 1.0, making it almost a mirror image of the chart above. Chances are good that the same threads would shop up as loose from the pack.

Whichever scale is used, the results would immediately show who has moved the furthest away from the center in relative terms. It would immediately and obviously show the biggest movements: those like Gateway that started high and went low, and those who like Nortel started low and went high. In this respect, it would allow us to compare companies directly, and it would quickly spotlight the most unusual changes.

Go to Part Two