We note that while Vehicle C failed just the uphill test, Vehicle B apparently failed multiple on-road tests (though not as badly as Vehicle A). We note this because in October 2014, the ICCT performed a similar round of tests on 15 European diesel cars, comparing the results of Euro 6 standard lab tests to road tests using portable emissions measurement systems. Out of 15 cars tested, 14 were over the limit when tested on the road using portable equipment, while all 15 were under the limit when tested in the lab.

The ICCT and WVU reported the discrepancy to both CARB and the U.S. Environmental Protection Agency, which began an investigation. We now know that Vehicle A, which the study identified only as a 2.0-liter diesel engine passenger car that used a NOx abatement technology called "lean-NOx trap," was in fact a Volkswagen model. And that subsequent investigation found that the car's control software was programmed to turn up the lean-NOx trap when the car was being tested in the lab, and then to turn it back down when the car was actually being driven on the road.

It makes perfect sense, when you think about it. Many emissions control systems on both cars and trucks reduce the amount of power delivered to the wheels or increase the amount of fuel needed to turn them. But during a typical emissions test, the weight of the vehicle doesn't have to be pulled uphill, and performance isn't an issue. Clean air is what's needed, not horsepower.

However, as we just noted, two out of the three, and then 14 out of the 15 diesel cars examined, failed the real world tests but passed the artificial tests. We doubt they were all Volkswagens. So we doubt this is the end of the story. This could be just the beginning of a widening scandal that eventually ensnares multiple manufacturers of allegedly clean diesel engines.

The Clean Air Act

The Clean Air Act, first passed by the U.S. Congress in 1963 and amended four times since, is an environmental law that gives the EPA the power to initiate a recall of vehicles whose emissions exceed defined limits. Vehicle manufacturers are required to have their products tested to make sure their emissions are below those limits. And they are required to warrant their emission control systems for much longer periods than are typically offered for the rest of the vehicle's components.

Volkswagen's passenger cars equipped with diesel engines passed these federal tests, as well as numerous individual state tests required to keep the vehicles registered and on the road. But now the EPA accuses VW of installing specific software that instructed the engines to reduce emissions whenever they detected that a test was under way. When the tests were over, the software returned operations to normal, and emissions increased far above the EPA's limits.

In a letter the EPA sent to Volkswagen on Sept. 18, the EPA outlined the two Clean Air Act violations it alleges. Specifically, 1) "the EPA has determined that VW manufactured and installed defeat devices in certain model year 2009 through 2015 diesel light-duty vehicles equipped with 2.0 liter engines. These defeat devices bypass, defeat, or render inoperative elements of the vehicles' emission control system that exist to comply with CAA emission standards. Therefore, VW violated section 203(a)(3)(B) of the CAA, 42 U.S.C. section 7522(a)(3)(B)."

And then 2), "the EPA has determined that, due to the existence of the defeat devices in these vehicles, these vehicles do not conform in all material respects to their vehicle specifications described in the applications for the certificates of conformity that purportedly cover them. Therefore, VW also violated section 203(a)(1) of the CAA, 42 U.S.C. section 7522(a)(1), by selling, offering for sale, introducing into commerce, delivering for introduction into commerce, or importing these vehicles, or for causing any of the foregoing acts."

Admitting Defeat Device Usage?

Amazingly, it seems that Volkswagen has already admitted it used a defeat device to pass federal and state emissions tests. In a letter sent by CARB to Volkswagen, the state regulators outlined their case, and noted the company's response:

"As you know, these discussions over several months culminated in VW's admission in early September 2015 that it has, since model year 2009, employed a defeat device to circumvent CARB and the EPA emission test procedures."

How did the defeat device work? Back to the EPA letter: "VW manufactured and installed software in the electronic control module (ECM) of these vehicles that sensed when the vehicle was being tested for compliance with EPA emission standards."

It was basically a software switch that determined whether the car was on the road, or when it was undergoing an emissions test, by monitoring the position of the steering wheel and the speed of the engine. If the engine is idling, the wheels are still, and the steering wheel is stationary, the car is probably stopped at a red light. But if the engine is revving and the wheels are spinning but the steering wheel is still, it's probably being tested (or it's being driven down an incredibly straight highway).

"During EPA emission testing, the vehicles' ECM ran software which produced compliant emission results under an ECM calibration that VW referred to as the "dyno calibration" (referring to the equipment used in emissions testing, called a dynamometer). At all other times during normal vehicle operation, the switch was activated and the vehicle ECM software ran a separate "road calibration" which reduced the effectiveness of the emission control system (specifically the selective catalytic reduction of the lean NOx trap). As a result, emissions of NOx increased by a factor of 10 to 40 times above the EPA compliant levels, depending on the type of drive cycle (e.g. city, highway)," the letter stated.

In a press release the EPA issued on September 18, the agency said it was pursuing Volkswagen as a matter of public health. "Using a defeat device in cars to evade clean air standards is illegal and a threat to public health," said Cynthia Giles, assistant administrator in the EPA's Office of Enforcement and Compliance Assurance.

"Our goal now is to ensure that the affected cars are brought into compliance, to dig more deeply into the extent and implications of Volkswagen's efforts to cheat on clean air rules, and to take appropriate further action," added California Air Resources Board Executive Officer Richard Corey in another section of the press release.

However, the EPA also pointed out that for owners, these cars are safe and legal to drive. The agency will not confiscate any vehicles or require owners to stop driving them. Also, the agency said that Volkswagen will be required to implement corrective action at no cost to the owner.

Almost Half a Million Vehicles

The EPA said the allegations cover roughly 482,000 diesel passenger cars sold in the United States since 2008. Affected models include:

• Audi A3 (model years 2010-2015)
• VW Beetle (2012-2015)
• VW Beetle Convertible (2012-2015)
• VW Golf (2010-2015)
• VW Golf Sportwagen (2015 only)
• VW Jetta (2009-2015)
• VW Jetta Sportwagen (2009-2014)
• VW Passat (2012-2015)

After the EPA and CARB made their announcements, Volkswagen responded. "I personally am deeply sorry that we have broken the trust of our customers and the public," stated now-former-CEO Martin Winterkorn in a prepared statement released on September 20. "We will cooperate fully with the responsible agencies, with transparency and urgency, to clearly, openly, and completely establish all of the facts of this case. Volkswagen has ordered an external investigation of this matter."

Earlier today, the company's board announced that the law firm of Jones Day is going to conduct the external investigation. And it said that technical solutions to the emissions problem are being developed and will be presented to responsible authorities before the end of October. Also, it said that the company's initial evaluation estimates that some five million vehicles out of the 11 million worldwide total will require "a service procedure" to correct the emission problem.

Another Diesel Test Failure

Has there ever been a case like this before? It's not exactly the same, but back in 2010, Navistar International Corp. faced a similar dilemma. The EPA had ruled in 2001 that the diesel engines used in all highway trucks had to reduce their NOx emissions levels by 95%, to a level below 0.2 grams of NOx per horsepower per hour. But it also gave engine manufacturers nine years to reach that goal.

Cummins, Caterpillar, Detroit Diesel, Volvo Trucks, and other truck engine manufacturers designed new engines that met the stricter standards, using a type of technology they called selective catalytic reduction. But Navistar's engineers pursued a different type of emissions reduction technology that it called advanced exhaust gas recirculation, and they couldn't design a compliant engine in time.

Fortunately, section 7525(g)(1) of the Clean Air Act allows heavy-duty vehicle engine manufacturers to get a certificate of conformity even if they fail to meet the standards, as long as they pay a nonconformance penalty. At first, Navistar "paid" the EPA by turning in credits it had accumulated in years past for the sale of engines that were cleaner than the standards in force at that time. But by early 2012 those credits had run out.

Eventually, Navistar's competitors sued the EPA, essentially complaining that since they had all met the new emission standard by 2010, Navistar couldn't continue to claim it was too difficult to do so. But the lawyers attacked the penalty regime on a technicality: the alleged lack of public notice before the rulemaking. They prevailed, and Navistar switched to selective catalytic reduction technology.

In other words, at some point, Volkswagen and Navistar faced a similar choice. Their engineers needed to design a diesel engine that would pass the EPA's stricter emissions tests. VW chose lean-NOx trap technology. Navistar chose advanced exhaust gas recirculation technology. Neither technology resulted in designs that would pass the test.

Navistar decided to pay the penalty for failing the test. VW decided to cheat on the test. Navistar decided to switch to Cummins engines, buying itself some time until it could develop its own selective catalytic reduction engines that could pass the EPA test, which it finally did in 2013. In contrast, VW tried to get away with it, until some precocious college students ruined everything.

Implied Warranty Exposure

Paul Wojcicki, a warranty lawyer with decades of experience representing vehicle manufacturers in class actions, product liability claims, and warranty and service contract actions, is a shareholder in the law firm of Segal McCambridge Singer & Mahoney Ltd. He said he is astonished at what he's hearing about the VW scandal, specifically the admission that a defeat device was used to pass the emissions tests.

Wojcicki said if the company really did overtly design its product to fool the tests, there are probably three types of warranties involved. First is the implied warranty of merchantability, which asks if the vehicle is reasonably fit for its intended purpose. "VW has been quick to point out that it doesn't affect safety or drivability, but I think merchantability goes a little beyond that," he said.

Second is the emissions warranty. "You're expressly warranting that your vehicle is designed in such a way that it complies with federal emission standards," he said. "So that warranty has clearly -- admittedly, given VW's statements -- has been breached."

In this case, nobody can directly link a Volkswagen vehicle to a death or injury caused by excess nitrogen oxide levels in the air. However, nitrogen oxides can also have serious adverse health effects. They're a precursor to ground-level ozone, which causes significant respiratory problems that can lead to premature death.

And third is the performance warranty. When driven, these diesel cars are emitting more pollution than is allowable, because the emissions control system has been turned down. Turn it back up, and performance suffers. Poor acceleration would result. But that's a design defect, not a repairable mechanical breakdown, so the warranty work won't be as simple as changing a filter or two.

Incredibly, the remedies available under the implied warranty, which comes out of the Uniform Commercial Code, could cost VW the most. The best-case scenario is they somehow fix the cars so their emissions are under the limit. Another possibility could be engine replacements, which would be expensive, but not cataclysmic. "And the upper end would be they have to buy back 11 million vehicles worldwide, or at least 482,000 in the U.S.," he said.

Wojcicki suggested comparing this VW scandal to the Ford Pinto episode back in the 1970s, which Malcolm Gladwell recapped comprehensively in a May 4 piece for the New Yorker magazine.

Back then, Ford essentially looked at the fuel tank problem from a cost perspective, finding that from a frequency and severity perspective that it wasn't worth correcting, even though people died. And then when it became a public scandal, it took a big hit in terms of reputation.

"But this is something different," he said. "This involves intentional conduct, and it would be treated differently than merely negligent conduct. There was an intent to defeat the regulation. These guys tried to subvert the standard. I've never seen anything come out of this nature. This is potentially more devastating."

Potential Warranty Liabilities

Among warranty professionals, the worst-case scenario is that everything your company sold comes back broken, and needs to be repaired or replaced. In other words, it's a 100% failure rate. How much will that cost?

From the EPA's press release, we know the frequency. There are 482,000 VW and Audi cars sold in the U.S. subject to some sort of emissions recall, to make them compliant with the Clean Air Act. It could be done by the installation of some sort of truck-like selective catalytic reduction apparatus on existing diesel engines that somehow reduces the NOx emissions and captures the soot. Or it could be as simple as replacing the electronic control module software.

Or perhaps it can't be done? Perhaps the operation of the filter requires the diversion of so much horsepower from the wheels that it turns a Beetle into a Trabant? Perhaps the real reason VW's software reduced emissions controls once the test was over was because they had an unacceptable impact on power and performance when they were turned on?

So perhaps the worst-case scenario for VW really is a complete engine replacement in 482,000 vehicles. What would that cost?

To find out, we did a bit of shopping. An online parts distributor called Car Monkeys is selling a used 2010 Volkswagen Golf 2.0-liter diesel engine assembly for $5,560. Another online parts site, Automotix LLC, has 28 different offers for used 2.0-liter engines for a 2012 Volkswagen Beetle, ranging in price from $2,057 to $2,732. At LKQ Online, a 2014 Volkswagen Jetta used 2.0-liter diesel engine assembly is priced at $2,100.

So let's pick a round number upper-bound price of $2,700 for the street price of a used late model Volkswagen replacement engine. The cost of 482,000 engine replacements on the open market would therefore be $1.3 billion, just for parts. If we assume an additional $1,000 per vehicle for independent repair shop labor, we're up to $1.8 billion. You can adjust these estimates on your own for new OEM parts and dealer labor rates. And you can adjust the estimates up or down based on whether you think the replacements will run on diesel or gasoline.

That's our estimate of the maximum warranty cost for the U.S. portion of the installed base: $1.8 billion. And that's actually quite significantly smaller than the $2.5 billion allocation General Motors made in the first half of 2014 for its massive ignition switch recall, which had a much larger frequency but a much smaller severity than VW's current problem.

The diesel engine recall is expected to begin within a year, after VW has outlined the corrective actions it will take and has developed a recall plan. And then, once the EPA has approved the plan, VW and Audi will begin issuing recall notices, and affected vehicle owners will be notified of that recall.

But then there are another 10.5 million allegedly non-compliant VW 2.0-liter diesel engines on the roads of other countries. Will they also be subject to a recall and an engine repair or replacement? Multiply that $3,700 cost by 10.5 million additional vehicles and you have to spend another $38.9 billion for parts and labor worldwide. That's a payout very few companies could survive.

We'll leave that question to the individual country regulators to decide. First of all, not all national regulators apparently hate dirty diesel engines, because some of their cities are filled with them. The U.S. regulations are comparatively strict. Second of all, there are national political questions to be answered: do they really want to drive one of Europe's largest manufacturers out of business all because of some dirty air?

Volkswagen's Warranty Costs

Among the major automakers worldwide, Volkswagen has consistently had relatively high warranty expenses, as we most recently documented in the July 9, 2015 edition of Warranty Week. As the chart below shows, the percentage of the company's worldwide automotive revenue that's spent on warranty claims has ranged from a low of 3.5% in 2012 to a high of 4.7% in 2009, while the amount it sets aside as warranty accruals each year has ranged from 4.5% to 5.8% over the past 12 years.

Competitors such as Daimler, BMW, Fiat, and PSA Peugeot Citroen have warranty expense rates below that level. But more importantly, some such as BMW and Daimler have reduced their expense rates over time.

VW's claims payments per year have ranged from 3.5 billion euro in 2003 and 2004 to well over 7 billion euro in 2013 and 2014. So even if they hit the max and end up spending 1.6 billion euro on 482,000 engine replacements, it won't double or triple their worldwide warranty costs. If they have to replace five million engines, well, that's another story.

However, there's another way to look at VW's warranty expenses. Last year, VW sold 10.2 million vehicles worldwide, and set aside 9.7 billion euro in warranty accruals. Divide one figure by the other, and convert it into dollars, and you have $1,264 per vehicle. Do that for each of the other 11 years for which we have data, and you have the chart below.

As we detailed in the July 9 newsletter, VW isn't setting aside the most per vehicle. That undesirable title goes to either BMW or Daimler, in every year except in 2006 (though VW placed a close second in 2008 and 2013). However, BMW and Daimler are making millions of luxury vehicles with much higher average selling prices. Volkswagen is the "people car," selling millions of mid-priced Jettas, Golfs, and Beetles. That's why the amount of accruals per vehicle is such a comparatively high percentage of VW's revenue: the revenue per vehicle is much lower.

Now we don't know how much VW accrues per vehicle sold in the U.S., as opposed to in Europe or in China. It could be higher, or it could be lower than $1,264 per vehicle. All we have is the worldwide average. But there's no way around it: an additional $3,700 per vehicle is going to obliterate the company's loss triangles. We're talking about diesel model warranty costs that could be as much as triple what the actuaries expected. That's going to be quite a big change of estimate.

The Penalty Phase

And then there are the penalties to be paid. The EPA has two separate penalties: For installing the defeat device, and therefore violating section 203(a)(3)(B) of the Clean Air Act, it can assess up to $3,750 per violation. For selling the cars without an EPA-issued certificate of conformity, and therefore violating section 203(a)(1) of the Clean Air Act, it can assess an additional penalty of up to $37,500 per violation.

We're not lawyers, and we're not accountants or regulators either. So nothing we say should be interpreted as legal or financial advice. But aren't the cars separate from the certificates? Each car pollutes the air individually, but each certificate covers multiple vehicles in what the EPA calls a "test group." So they couldn't assess the larger fine on a per-vehicle basis. Could they?

Let's assume the EPA assesses just the smaller penalty on a per-vehicle basis, and the larger penalty on a per-model-per-year basis. For the Audi, that's six model years and six certificates of conformity obtained fraudulently. For the Passat, that's four model years and four certificates of conformity obtained fraudulently. And so on. We have 38 model years spread across eight models.

In other words, let's assume VW pays 482,000 penalties of $3,750 each for the dirty engines, and 38 penalties of $37,500 each for the non-conforming test groups. That's a total fine of just over $1.8 billion.

And then there are the state fines. California has been deeply involved in this investigation, but other states also test cars for emissions, and they also might feel injured by VW's ruse, especially if there are six, seven or eight zeros attached to the settlement figures.

Such fines would not be unprecedented. Remember, we're talking about the EPA and the U.S. Department of Justice enforcing anti-pollution statutes, not the FTC, FDA, NHTSA, or some other federal agency enforcing product safety standards. The record in this regard is the $18.7 billion environmental fine levied against BP plc after the 2010 Deepwater Horizon oil spill, which in turn was part of a $54 billion package of clean-up costs, damage settlements and penalties.

The National Highway Traffic Safety Administration assessed two fines of $35 million each against Honda Motors for failing to report deaths, injuries, and certain warranty claims to the federal government in violation of the TREAD Act. The first fine was a result of Honda's failure to report 1,729 death and injury claims to NHTSA between 2003 and 2014. The second was caused by the automaker's alleged failure to report certain warranty claims and claims under customer satisfaction campaigns during the same years.

The defective Takata airbags recalled by numerous auto manufacturers have a total cost now projected at $2.1 billion, based on 34 million replacements. And that doesn't include the cost of settling all the lawsuits. But that's mostly a warranty and recall expense borne by the manufacturers, not a penalty paid to the government.

Nevertheless, NHTSA said 2014 was a record year for civil penalties. Last year, NHTSA fined General Motors $35 million for the ignition switch defect, which eventually expanded into a recall of nearly 2.2 million cars.

NHTSA also fined Hyundai Motor America $17.35 million for the failure to issue a recall in a timely manner, and fined Ferrari $3.5 million failing to submit early warning reports. But the maximum amount it can assess in any single fine remains capped at $35 million, so don't expect any nine-figure settlements from that agency.

And while the idea of fining VW once per model per year for the fraudulently obtained certificates of conformity may seem a bit absurd, it's based upon our observation of the way that NHTSA gingerly assessed the TREAD Act fines against Honda. Yes, they were a record $35 million each, but they could have been far larger in total. Manufacturers are required to submit the early warning reports on a quarterly basis. That means four reports a year. So in a worst-case scenario, Honda could have been hit with four pair of fines per year from 2003 to 2014, rather than one pair of fines covering scores of separate filings over twelve years in total.

Pay the Piper

But there's yet another penalty that will never be paid. European and American regulators either knowingly or negligently designed emissions tests that could be fooled, sometimes with helpful input from the manufacturers themselves. And while researchers have been documenting the gap between lab test and real world results for quite some time, this is the first time it's become such major news. Is it possible this is the first time anyone realized that allegedly clean diesel vehicles were dirtier than they should be?

Meanwhile, at least five million and maybe as many as 11 million non-compliant VW diesel vehicles are still on the roads, and will be until a fix can be designed for them. And there are other diesel cars made by other companies that also failed the ICCT and WVU tests in real-world settings, though not by as much as Volkswagen. We understand why the testers wouldn't want to name names, given how libel laws work and how little money poor students and nonprofit organizations generally have to spend on legal representation. But why not announce the names of vehicles that actually did pass the tests both in the lab and on the road?

Nevertheless, there are still all those older trucks and buses that were never tested but are nevertheless still belching soot and other pollutants. If vehicle owners and drivers dislike the way the anti-pollution technology compromises performance, wouldn't they do everything they can to extend the life of their old engines?

Who will pay a penalty when there's another ozone alert in a major world city? Who will pay the medical bills of the hundreds of millions of asthma sufferers? It certainly won't be the vehicle manufacturers that put their diesel cars, trucks and buses on the road, spurred on by policymakers who believed in clean diesel. And it won't be the regulators that helped them pretend that their products were clean just because they passed their shallow tests.