A rotten apple spoils the barrel, as the old saying goes, and because of regulatory politics and the incentives of agency officials, the spoilage often proceeds unimpeded even if the rotten one and the others occupy separate barrels. Witness for example the Volkswagen diesel emissions scandal that emerged in 2015, a truly brazen act of business malfeasance that will cost VW at least $15 billion in fines and other costs.
Like all internal combustion engines, diesels produce a range of effluents, with nitrogen oxides (NOx) the central one of interest in the context of the discussion here. To a degree greater than that for gasoline engines, the design of diesel engines must confront the problem of reducing emissions of both NOx and fine particulate matter (PM2.5). Lower engine-combustion temperatures reduce emissions of NOx but increase emissions of PM2.5, while higher engine-combustion temperatures increase NOx emissions and reduce those of PM2.5.
Until 1999, US emissions rules allowed higher NOx emissions for diesel engines than for gasoline engines. Accordingly, for years the optimizing solution for the engine manufacturers was higher engine-combustion temperatures and a higher level of NOx emissions. But a regulatory change (“Tier 1” and “Tier 2”) in 1999 required both diesel and gasoline engines to meet the same NOx emissions standards, making necessary the installation of additional emissions control equipment.
Two NOx control technologies are both practical and economic for diesel engines. A selective catalytic reduction (SCR) system adds urea and water (“diesel emissions fluid”) to the exhaust flow, which breaks the NOx down into nitrogen and carbon dioxide. This system requires a reservoir for the diesel emissions fluid, a pump, and a delivery system for the exhaust flow.
A nitrogen oxide trap “absorbs” nitrogen oxides during the fuel combustion process; when the trap is saturated to capacity, some diesel fuel is injected into the trap, converting the NOx into nitrogen and regenerating the trap system. Crucially, this use of diesel fuel does not power the engine, so fuel economy is reduced along with engine performance.
A Volkswagen car is pictured during a test at a technical center. REUTERS/Dado Ruvic
Unlike the rest of the industry, which adopted the SCR system for its diesel vehicles, VW opted for the nitrogen oxide trap, apparently in part because the space requirements for the SCR reservoir, pump, and delivery system would interfere with the need for other equipment demanded by many consumers, particularly in the small passenger cars incorporating the VW diesel engines. One trade publication noted that
There’s more to putting a urea tank on a car than just lashing it down with zip ties. The floorpan will likely change, the addition of a secondary filler can mean retooling a quarter-panel, and the reengineered car has to go through full crash certification. Golf V owners liked their multilink rear suspensions, a feature that probably would have to be scrapped in favor of a more compact torsion beam to leave room for the tank. (The larger Passat got a urea system in 2012, but VW programmed it so that, apart from the emissions test, it would be stingier with the urea injections, meaning that the owner wouldn’t be inconvenienced with refilling the tank as often.)
With Tier 2 looming, VW faced spending millions on an aging product to make its diesel engines legal in the U.S., one of the smallest diesel-passenger-car markets in the world. In 2007, diesel passenger cars represented only about 0.2 percent of the American market. Factor in the contemporaneous collapse of the world economy and the plunge in U.S. vehicle sales, and VW’s engineers were painted into a corner.
Moreover, several observers have noted that the VW diesel systems were designed for Europe, where the VW diesel market is much larger than in the US, and where the NOx emissions standards are significantly less stringent. This meant that the investment costs just noted were unlikely to have been justifiable in a standard business case (or benefit/cost) sense. In any event, VW advertised its “clean diesel” technology as one that “reduces nitrogen oxide (NOx) emissions by up to 90 percent by making internal engine modifications and implementing a NOx storage catalytic converter,” offering “fewer NOx emissions than comparable gasoline engines,” and meeting “the strictest EPA standards in the U.S.”
Those promises at best proved very difficult to achieve: Consumer Reports found a “noticeable decline in fuel economy” for the VW models when the nitrogen oxide trap systems were operating properly. Nonetheless, VW certified its compliance with the Tier 2 NOx standards, and the EPA issued a Certificate of Conformity allowing the sale of its diesel passenger cars in the US
A team of researchers at West Virginia University conducted a study (undated, but apparently published at least in draft form in May 2014) of on-road emissions from diesel vehicles, under a contract from the International Council on Clean Transportation. The researchers found that emissions levels for some vehicles exceeded US certification standards, with some of the differences very substantial; subsequent vehicle testing by the EPA found that VW NOx emissions were up to 40 times greater than the standard for its 2.0 liter diesel engines for model years 2009–15; and nine times the standard for the 3.0 liter engines for model years 2014–16. The disparity between the emissions standard, as certified by detailed vehicle tests on dynamometers, and actual on-road performance is more complex than commonly assumed, as discussed briefly below.
VW asserted initially that the disparity was the result of a software problem that could be addressed with a voluntary recall. EPA then discovered that VW had installed software in the vehicles’ computer systems that sensed when the vehicles were being tested, activating the nitrogen oxide trap systems. When not being tested—when being driven on the road—the software simply shut down the traps, thus “defeating” the emission control systems.
It is obvious that the goal was to certify compliance with US NOx emissions requirements while avoiding the costs and performance degradation of meeting them in on-road performance for vehicles designed for the European market and sold in a smaller US market for diesel passenger autos.
Note that VW is the only auto manufacturer accused of explicitly installing such systems to defeat the NOx emissions control systems. In the other recent cases (say, since 2000) of the use of defeat devices or technology—Harley-Davidson (2016), Casper’s Electronics (2013), and Edge Products (2013)—all were aftermarket producers rather than the vehicle manufacturers themselves. Other important “defeat” cases are older—1998 in particular—but for the most part involved manufacturers of heavy truck engines. Those cases were settled with consent decrees and without admission of illegality, unlike the VW case. It is worth summarizing in a simple table the differences between the Volkswagen NOx case and the practices of the rest of the industry.
Diesel NOx Emissions: Volkswagen vs Other Auto Manufacturers
| Volkswagen | Other Manufacturers | |
|---|---|---|
| Emissions Technology | Nitrogen Oxide Trap | Selective Catalytic Reduction |
| Test Performance | Trap Controls NOx | SCR Controls NOx |
| On-Road Performance | Software defeats NOx controls | SCR Controls NOx |
To say that the on-road emissions of a given vehicle exceed the allowed emissions standard or the emissions measured during dynamometer tests does not mean necessarily that “cheating” is the reason. In the VW case, the explicit effort to defeat the emissions control equipment was an obvious and deliberate evasion of the relevant emissions regulations. Consider instead a recent case involving Hyundai and Kia, settled by consent decree in November 2014; it resolved allegations that its fuel economy estimates were biased upward, as well as alleged violations of the Clean Air Act created by sales of more than one million vehicles that will emit about 4.75 million metric tons of greenhouse gases in excess of the emissions that the manufacturers certified to the EPA.
Apart from the fact that Hyundai and Kia deny violations of the law, unlike the case for VW, one obvious problem is that the specific protocols of the complex dynamometer tests, however honestly and diligently designed, may not replicate actual on-road driving behavior and conditions by vehicle owners and operators. A large number of variables affect emissions performance, and it is extremely difficult to choose among the myriad sets of relevant conditions. Accordingly, emissions observed as greater in on-road conditions than those measured in dynamometer tests may substantially or wholly result from the assumptions and biases inherent in the latter. That is the clear implication of the findings in the West Virginia University study (Figures 4.15–4.33) noted above. It is worth repeating: That disparity can be observed even if the emissions control equipment is operating as designed, as certified to the EPA, and as advertised to the public.
Nonetheless, the VW scandal reportedly has yielded a series of other investigations. Perhaps actual cheating will be discovered, although no such evidence yet has been made public. And there is a case to be made that some due diligence on the part of US regulators is appropriate, in that an absence of such effort could be predicted to yield an erosion of the emissions regulatory framework, and environmental quality more centrally.
U.S. Rep. Diana Degette (D-CO) holds a diagram of an automobile’s exhaust system while questioning Volkswagen North America CEO Michael Horn during a House Energy and Commerce Committee hearing investigating the company’s admission it cheated U.S. emissions tests, on Capitol Hill in Washington October 8, 2015. REUTERS/Joshua Roberts
But it is worth bearing in mind that the regulatory bureaucracy itself has incentives that are not entirely salutary, as the VW cheating scandal has provided a rationale for what can be described as fishing expeditions in pursuit of objectives not at all environmental. The investigations provide an argument for increases in the regulatory agencies’ budgets, always a hidden but central goal of government bureaus. The fines and settlement payments, even if given entirely to the US Treasury, are likely to increase the agencies’ budgets indirectly as Congress shares the proceeds with the agencies implicitly as part of the process of bargaining over the budget. The fines and settlements can further the ideological goals of the regulators, as the manufacturers are forced to fund particular projects not approved by Congress in the budget process, a perverse system that Attorney General Jeff Sessions has begun to constrain.
There is the further matter that the Trump administration has reinstituted the mid-term review of the corporate average fuel economy standards for model years 2022–25, after the Obama administration had canceled that review just before leaving office. It is at least plausible that the VW cheating scandal has given the federal regulators a bargaining chip with which to induce the manufacturers to agree to a CAFE standard higher rather than lower.
In short, it is perverse to fail to distinguish between the behavior of VW and that of the rest of the industry. Such a distinction is easy to miss in the public discussion, but policymakers must keep it very much in mind.