If you begin from an assumption of consumer rationality, you'll prefer a carbon tax or some form of emission trading. Announce today a schedule of Pigovean carbon taxes and how they will affect petrol prices as they ramp up over time. Then let customers decide how to re-optimise when buying cars. We'd expect an increase in demand for cars with better fuel economy. Cars with worse fuel economy will start having to sell at a discount. Manufacturers adjust their product mix to account for changing demand and aggregate fleet composition changes over the longer term.
If you start from an assumption of consumer stupidity, you'll prefer regulations targeting car manufacturers mandating fuel economy standards. If car buyers are myopic and stupid, they'll fail to account for the higher lifetime cost of a car with worse fuel economy. Because customers are stupid in this way, manufacturers will not adjust their product mix to shift towards cars with better fuel economy - there's no change in demand for more efficient cars even with a well-publicised schedule of future tax increases. And so direct regulation has to be used. There are problems with this and lots of them - all the gaming of US CAFE standards and redefinitions of what constitutes a truck as most obvious example. But it could be a second best if car buyers are really stupid. Or, if they're just really really short-sighted. We'll also have to assume that car buyers do not change their behaviour by a lot when the price of driving a kilometer goes down by a lot.
What happens if we look to the data? Busse et al in the latest American Economic Review find pretty good evidence that car buyers' demand for fuel economy is sensitive to petrol prices. They conclude (ungated versions):
We estimated that a $1 increase in the price of gasoline increases the market share of cars in the highest fuel economy quartile by 21.1 percent and decreases the market share of cars in the lowest fuel economy quartile by 27.1 percent. We also estimated the effect of a $1 increase in gasoline prices on unit sales of new cars and found that sales in the highest fuel economy quartile increased by 10–12 percent, while sales in the lowest fuel economy quartile fell by 27–28 percent. We estimated the effect of gasoline prices on the equilibrium prices of new cars and found that a $1 increase in the price of gasoline is associated with an increase of $354 in the average price of the highest fuel economy quartile of cars relative to that of the lowest fuel economy quartile. For used cars, the estimated relative price difference is $1,945.
We used these estimates to investigate whether the changes in equilibrium prices for new and used cars associated with changes in gasoline prices show evidence that consumers undervalue future gasoline costs of cars with different fuel economies relative to the prices of those cars. This could be thought of as a necessary condition for effective policy: the more car buyers discount future fuel costs, the less effective a gasoline tax or carbon tax will be in influencing vehicle choice. Using several different assumptions about vehicle miles traveled, a range of assumptions about the elasticity of demand, and comparing the relative price differences between different quartiles, we find little evidence of consumer myopia. Many of our implicit discount rates are near zero; most are less than 20 percent.So the "people aren't stupid and weigh costs over time in a sensible fashion" model seems the better baseline approach.
Now imagine that you set a fuel economy standard instead of a carbon/petrol tax in a world where customers are forward-looking and not idiots. Well, once they've bought the more efficient car, the value they derive from burning another litre of petrol increases substantially: they can drive farther, and they're not charged any more for that litre of petrol. And so a lot of the reductions in carbon emission you might have expected get whittled away by that people drive more. If you'd done it instead with a petrol tax, the marginal cost of another litre of petrol is higher. People still flip to the more efficient vehicle, but petrol usage doesn't rebound as much as consequence because the marginal cost of a litre is higher.