Daniel Stern's website is not lighting gospel and references studies performed four or more decades ago (not to completely discount the work done in those studies). Furthermore, one of Stern's primary arguments is that filtering a halogen or incandescent light source to make it selective yellow reduces its intensity is valid and true. However, when discussing HID lamps, their color temperatures are a result of the chemical makeup within the arc tube and it's possible to have equal brightness across a range of color temperatures. For the purposes of this discussion, we are talking about variations within the "white" spectrum - cool white vs warm white - not selective yellow.
There are numerous studies on roadway lighting with the introduction of LED streetlights that make some very interesting conclusions on roadway lighting (both fixed street lights and car headlights). In general, cool white and daylight is not preferred for a variety of reasons ranging from light pollution (mostly irrelevant when talking about headlights) to glare to the differences in how the color of a light source affects a person's scotopic vs photopic vision. Stern addresses some of the issues in the link you provide. Rayleigh scattering or otherwise, bluer light performs more poorly in automotive lighting applications than warmer light (within the range of "white" light).