Before fully investigating to what extent excess vehicle emissions had a negative effect upon health or the environment in Canada or whether a I/M program would be an effective strategy for reducing emissions, Environment Canada formed the CCME I/M Working Group to develop the ``best'' or ``most effective'' I/M program. Most of the members of the Working Group had a clear vested interest in establishing I/M programs in Canada (see figure 9). Other members of the Working Group had no opportunity to acquire any independent knowledge of I/M programs and therefore had to rely almost entirely on CCME background material for answers to technical and scientific questions. Environment Canada apparently found no need for input from independent scientists who could read technical reports, understand the statistics, and help guide the others.

Formal meetings of the Working Group were held in 1994 to write the original Code of Practice and again in 1997 to make revisions. Environment Canada commissioned two background documents, one for the original Code and the second for the revised code. Both were written by the same consulting firm. Responding to an APA report claiming that the Environment Canada background document for the original code was biased and promoted I/M uncritically, Douglas Cope, the consultant, wrote: ``Of course the draft Background Report, produced by D. Cope Enterprises, favored I/M, this was what my firm was contracted to do!''33

In his letter, Mr Cope wrote that Environment Canada and the CCME had decided in favour of I/M programs before the contract to develop the Code of Practice was let to his firm. The background report confirms that one purpose of developing a Code of Practice was clearly to promote I/M programs: ``The existence of a CCME Code of Practice will also assist provincial agencies in convincing both management and politicians that I/M should be implemented'' (CCME 1993: 2-8).

The manner in which the views and interests of both the non-governmental I/M lobby and various bureaucrats were promoted is indicated by table 5, which categorizes the list of references found in the Code, the revised Code, and the two background documents34 that were made available to working group members by Environment Canada.

Click Here to View Table 5

However, among the 9 ``other'' documents (which include a few newspaper clippings) there are two very interesting reports. The first is a report from the United States General Accounting Office (GAO) that describes a very serious problem:

[T]here are drawbacks to the [EPA's recommended] IM240 test procedure for both the identification and repair of out-of-compliance vehicles that could hamper the test procedure's effectiveness. Our review of EPA data found that over 25 percent of the vehicles that EPA tested using the IM240 test procedure failed an initial emissions test but passed a second emissions test, even though no repairs were made to the vehicles. These results raise questions about whether the IM240 test procedure is reliable in identifying out-of-compliance vehicles and whether inaccurate identification of emission problems could result in unnecessary repairs. (US GAO 1992: 1-2)

This report should have set alarm bells ringing when it was cited in the first background report, not only because it questions the validity of scheduled testing but also because the first CCME Code of Practice strongly recommended using the IM240 test procedure. Instead, this GAO report was not quoted but merely described as having ``suggested there were repeatability problems'' which were possibly ``not the fault of IM240;'' there followed a couple of sentences of exonerating speculation (CCME 1993: 5-17).

The second ``other'' CCME reference is to a peer-reviewed article published in 1987 by the Society of Automotive Engineers (SAE). Harold Haskew, the principal author of this article, is a well-respected expert in vehicle emissions. The article was listed among the references to the first background report but it does not appear to be cited anywhere in the text. It states:

Based on the in-use emission performance of today's [1987] closed loop systems and the cost of inspecting vehicles, it is difficult to support the need for an I/M program. If there are other justifications for an I/M program, a more meaningful approach would be to use a combination of a parameter [i.e., visual] inspection and the on-board diagnostic system. (Haskew et al. 1987)

If the interrogation of a modern vehicle's on-board diagnostic (OBD) system were made the primary feature of a modern emissions reduction program, it could cost less, be more convenient, and provide a better means of identifying vehicles with emissions problems than the dynamometer testing recommend by the CCME. However, this option was never seriously explored in the CCME material.

Environment Canada also overlooked a large number of scientific works that raise doubts about the effectiveness of its recommended I/M program and about I/M in general. For instance,

We compare roadside survey results at I/M and non-I/M locations in California, and observe identical tampering and overall failure rates at the two types of locations. We also show that motorists are taking steps to ``pass the test'' and that the high-emitting vehicles' idle-emissions performance and tampering rates in the roadside surveys are unaffected by the Smog Check test ... EPA's national tampering surveys also show little difference in tampering rates among areas with decentralized, centralized and no I/M programs in the United States. (Lawson et al. 1993: 1567)

Although the CCME strongly recommends the IM240 procedure, it does not accurately reflect reality:

Ironically, the IM240 [recommended by the CCME] is derived from Federal Urban Dynamometer Driving Schedule (FTP), which both the EPA and CARB [California Air Resources Board] have stated does not represent real-world driving and emission patterns.

... there is no solid scientific base to presume success of the enhanced I/M programs in reducing on-road emissions, the goal which they were originally supposed to meet. (Zhang et al. 1996: 1449)

Another study showed that Colorado's IM240 test produced a 12 percent average false high emitter identification rate (14 percent for 1991-1996 vehicles) and a 23 percent average false pass rate (Gallagher et al. 1997: 3-4).

I/M programs evaluated with the US EPA computer model often show impressive emission reductions, but when ambient air measurements are made, or other real-world means of measuring effectiveness such as remote sensing or roadside pull-overs are used, those reductions all but disappear, indicating a flawed computer model.

Unfortunately, our investigation and synthesis leaves us with little confidence in the ability of existing regulatory agency methodologies (including computer based models, and supporting data and analyses) to realistically and reliably assess and forecast performance, compare alternatives and provide a rational guide to policy. (Aroesty 1994: ix)

Ambient air measurements show that Minnesota's centralized I/M program provides no significant reduction of the target pollutant (carbon monoxide) (Scherrer and Kittelson 1994).

Although roadside remote sensing (RRS) has been summarily dismissed in the CCME background document, it is a technology that can cut identification costs by over 90 percent and has many other advantages over conventional I/M techniques.

Use of roadside remote sensing resulted in very low cost identification of high emitting vehicles (HEV). The HEV identification costs for this study were about $100 per vehicle; they would have been as low a $9 per vehicle if had we been able to pull over every passing HEV for repairs35 (Lawson et al. 1996b).