Bought a tiger, got a Hello Kitty? How to fix vehicle fuel economy fraud in China

China’s most powerful government-controlled television broadcaster, CCTV, is famous for sensational stories exposing product-quality issues. Its March 15 show made auto industry PR specialists in China nervous, as it turned its attention to a pretty special new fraud: vehicle fuel economy.

The broadcast alleged that official vehicle fuel consumption values, determined by testing over a standard driving cycle at government-certified testing facilities and then announced by the Ministry of Industry and Information Technology, can be manipulated in many ways during the test. And so the real fuel economy that consumers get can be significantly different from the vehicle label values. One subsequent news article concluded sardonically that “Consumers shouldn’t be in the position where they think that they bought a tiger, but received a Hello Kitty.” Meow!

The episode mentioned several ways that car manufacturers can manipulate fuel economy data during tests:

1. Customize their prototype vehicle for type-approval testing. Examples of this strategy include choosing the lowest-weight vehicle within each weight group, choosing tires with the lowest rolling resistance among the available tires, and using special computer calibrations customized for the best fuel-consumption results.
2. Change the road-load parameters to reduce the load placed on the vehicle during chassis dynamometer testing.
3. Manipulate the test environment, using higher test temperature to reduce emissions and slightly reduce fuel consumption.
4. Have the driver follow an optimized driving cycle during testing, to minimize acceleration and reduce fuel consumption.

This ability to manipulate the test results can be traced to three basic “flexibilities” in the test procedures for vehicle fuel consumption certification. First, it would be impracticable to require that manufacturers test every possible vehicle configuration, so the procedures allow manufacturers to group vehicles together and test one vehicle from each group. Second, the fuel-economy label values must be generated before production starts, so all certification testing is necessarily conducted on prototype vehicles. Third, because testing is necessarily done in a controlled environment on a dynamometer, in order that every vehicle can be tested in a similar manner, it is necessary to establish the proper aerodynamic and tire rolling resistance losses prior to the dynamometer testing (commonly referred to as “road load”). And, tire rolling resistance can have significant impact on vehicle fuel economy. Determining the proper road load is much more difficult and expensive than conducting dynamometer tests, so again vehicles are grouped together and a single road load test is conducted for each group. All of these flexibilities leave plenty of room for car manufacturers to game the prototype vehicle calibration and test weight, vehicle test temperature, and road load, and to use lower rolling resistance tires for testing.

The key to the accuracy of the fuel economy labels is ensuring that the prototype vehicles are representative of production vehicles and that the components installed for road load determination and dynamometer testing are representative of the average production vehicle. While there are technical fixes to the test procedures that can reduce gaming (for example, using smaller steps for test weight bins), the most effective fix is a fairly straightforward administrative one: conformity of production (COP) and in-use verification requirements.

COP refers to post-production testing conducted by the regulatory agencies in order to make sure the vehicles in mass production perform the same as their prototype models did on the certification tests. It can be applied not only to fuel consumption, but also more broadly to tailpipe emissions, safety, and other regulatory standards. To make such a program effective, vehicles subject to the COP tests must be randomly selected from the production assembly line, and the auto manufacturing plants must not receive advance notice. It is also important to break in the vehicle for at least 4,000 km prior to testing, to remove “green” engine and aftertreatment effects.

An in-use verification program is similar to COP, although it extends the requirements to ensure vehicles driving on the road are still compliant with the standards to which they were initially certified. This matters a lot for so-called criteria emissions (e.g., particulates, ozone, CO, NOx, SOx, lead), as emissions aftertreatment systems may deteriorate over time. The in-use requirement ensures that manufacturers build reliable products with guaranteed environmental performance over the vehicle’s useful lifetime (or established durability period). Fuel-efficiency technologies don’t deteriorate significantly, with proper maintenance, over the vehicle’s useful life. Therefore in-use verification matters less to vehicle efficiency, though it still has some merit.

The road load determination requires a separate post-production verification test. In this case, the in-use or COP vehicle is coasted down on a test track to verify the manufacturer’s pre-production road load. Note that a single in-use or COP vehicle can be used to verify both the road load and the dynamometer test values, or the verification programs can be conducted separately.

It is crucial that all post-production and in-use verification tests be done by government-owned test labs. Also, the penalties vehicles and manufacturers are subject to for non-compliance must be strong enough to deter further gaming, such as voiding product certification, product recall or civil penalties.

The Untied States Environmental Protection Agency (EPA) is the only government agency that has developed a comprehensive compliance program for vehicle emissions and fuel economy/CO2 emissions. The program includes compliance activities at various stages of vehicle design, production, and use: 1) Representative vehicle and tire selection during pre-production certification; 2) Confirmatory testing of about 10% of the manufacturers’ vehicles at EPA’s lab prior to issuing a certificate; 3) Selective enforcement audit, comparable to COP; 4) In-use surveillance performed by EPA; 5) In-use verification testing performed by manufacturers; 6) Detailed requirements for representative components and tires for road load testing; 7) Periodic confirmatory road load testing by EPA; 8) Recall; and 9) Warranties and defect reporting. The figure below shows a timeline and EPA/manufacturer responsibilities for these compliance activities. While most of the tests are done by manufacturers, EPA plays the key role of gatekeeper by conducting a few crucial verification tests.

In contrast, other countries, including China, have primarily focused on pre-production certification testing, and very little has been done on post-production and in-use stages for vehicle fuel consumption. Passenger car fuel-consumption regulations in China currently include a brief provision on production conformity; the regulatory agency can randomly sample a few vehicles from the production line for a verification test. The rule does not specify whether manufacturers are to receive prior notice; in practice, they very often do. And there is simply no in-use requirement for vehicle fuel consumption (or emissions).

Nor does the regulatory agency in China own the test labs. All tests are done by the so-called third-party certified test labs, which are often partially owned or shared by auto manufacturers. EPA used to allow certified third-party labs to conduct testing for imported “gray market” vehicles, but was forced to discontinue the program due to widespread cheating by the third-party labs.

China needs three things to protect its consumers and achieve the real-world fuel-saving benefits that are expected from the rules: purely neutral test labs, a real COP or in-use verification program, and a road load verification program.