Based on a review of available data, several conclusions can be drawn:

• IM programs continue to have the potential to significantly reduce emissions of CO, HC, NOX and particulate emissions from vehicles.

• As evidenced by the British Columbia program, to cite one well-studied example, these reductions really occur in a well-run program.

• Centralized programs are more likely to achieve substantial reductions because they tend to have better quality control, are easier to enforce, and can more cost effectively use more sophisticated test procedures. In Mexico City where both types of programs were run in parallel, it was concluded that decentralized programs were much more corrupt and ineffective.

• More advanced test procedures such as the steady state loaded test or a dynamic test tend not only to be much more effective in reducing emissions, but also when run in centralized, high volume facilities, they are at least twice as cost effective as in decentralized facilities.