There are several separate concepts to be addressed: fuel-efficiency, accident rate, fatality rate and experience in other countries.
Does vehicle speed affect the fuel-efficiency?
Fuel efficiency is measured in mpg or Litres/100 km - i.e. factoring out the speed travelled, and just looking at the distance you can get per volume of fuel. I am assuming that fuel-efficiency is a reasonable proxy for g(CO2). This ignores tire-wear, vehicle and road damage, change in transport consumption due to faster times, etc.
Fuel efficiency is generally reduced for high speeds e.g. higher than 55 mph (88 km/h). [Ref: Consumer Reports test]
Note: The Oak Ridge National Laboratory claims [Figure 4.2] that some modern cars have better fuel efficiency at 65 mph than 45 mph, but that (with occasional exceptions) 55 mph is still better - the slope points downwards for higher speeds. Wikipedia has a nice graph.
It seems reasonable to conclude that if the typical speed of cars is increased above 55 mph, average fuel efficiencies will drop.
Does vehicle speed affect the risk of accidents?
According to the World Health Organisation (WHO), YES.
WHO's World Report on Road Traffic Injury Prevention states that:
There is a large amount of evidence of a significant relationship between mean speed and crash risk.
Empirical evidence from speed studies in various
countries has shown that an increase of
1 km/h in mean traffic speed typically results
in a 3% increase in the incidence of injury
crashes (or an increase of 4–5% for fatal
crashes), and a decrease of 1 km/h in mean
traffic speed will result in a 3% decrease in
the incidence of injury crashes (or a decrease
of 4–5% for fatal crashes).
A meta-analysis of 36 studies on speed limit
changes showed, at levels above 50 km/h, a
decrease of 2% in the number of crashes for
every 1 km/h reduction in the average speed.
These claims are all referenced to the appropriate studies in the report.
For comparison, they also map excess speed to the equivalent Blood Alcohol Level equivalent.
Does vehicle speed affect the risk of accidents involving serious injuries and fatalities?
According to the World Health Organisation (WHO), YES.
Again, the WHO's World Report on Road Traffic Injury Prevention states that:
The probability of a crash involving an injury
is proportional to the square of the speed. The
probability of a serious crash is proportional
to the cube of the speed. The probability of
a fatal crash is related to the fourth power of
For car occupants, the severity of crash injury
depends on the change of speed during the
impact, usually denoted as Δv. As Δv increases
from about 20 km/h to 100 km/h, the probability
of fatal injuries increases from close to
zero to almost 100%.
They go on to discuss pedestrian survival rates when hit at high-speeds. Pedestrian accidents are, perhaps, unlikely on motorways, but the figures are quite appalling.
What is the German experience?
Germany, famously, has autobahns with areas where speed-limits are not enforced on some vehicles.
Some people have claimed that Germany does not have an increased accident-rate in areas with no blanket speed-limit. For example, Wikipedia's page on autobahns states:
The overall road traffic safety of German autobahns is comparable to and in some cases better than that of other European highways.
Moreover, international accident statistics demonstrate that limited access grade separated roads such as Autobahns and motorways have much greater road traffic safety regardless of speed limit, suggesting that high speed alone isn't a deciding factor.
I haven't found such evidence, yet. I hope someone can add to this answer.
The International Transport Forum's International Traffic Safety Data and Analysis Group (IRTAD?) has a database of figures by country. Looking at the IRTAD figures, Germany has a relatively low death rate by population and by vehicle km travelled. (Compare it to neighbours with similar terrain and climate.)
However, these figures don't separate out autobahn travel from other roads, and the unlimited sections of the Autobahn only account for a small percentage of the total travel. (Wikipedia claims "only two percent of the traffic in Germany runs on unlimited sections" without an explicit reference to support that.)
UPDATE: A fact sheet (non-peer-reviewed?) from the European Transport Safety Council (ETSC) sheds some light on this. Speed Fact Sheet #1, Feb 2008.
It points out that:
One argument used in favour of the status quo is that a similar proportion of deaths on the motorways in Germany occur on sections without speed limits as on sections with limits. According to the German Statistics Agency, of the 645 road deaths that occurred on motorways in 2006, 441 or 67% occurred on motorway sections without limits. This is consistent with the figures from the previous year, 2005, when 662 deaths occurred on motorways, 462 (70%) of them on unlimited sections (fig. 3):
But this similarity of percentages takes no account of traffic volumes on the different sections. Traffic volumes on sections without speed limits are likely to differ appreciably on average from those on sections with speed limits. This should be investigated to cast more light on the numbers of deaths on the two kinds of section.
Some graphs are provided that show, with limited statistical rigour, that a disproportionately high number of lethal accidents occur, per km of road, in unlimited areas. (To my eye, they seem to undermine the original claim of similar death rates for speed-limited and unlimited areas.)
I think it is clear from the evidence that, all else being equal, high-speeds significantly decrease fuel-efficiency, significantly increase the risk of accidents, and dramatically increase the risk of fatal accidents.
Whether this environment cost and cost in lives is worth it is not a scientific question. It is an environmental, economic, ethical and political one. If you wish to be allowed to speed anyway, fine, but don't expect the science to say it is without any cost.