Speed reduction is proposed to be a cost-efficient way to reduce emissions due to a non-linear relationship between fuel consumption and speed (approximated by a third power). 

Correlation between fuel prices and freight rates can lead to biased estimates. To disentangle the correlation between fuel prices and freight rates arising from changes in global demand, we instrument fuel prices on price changes triggered by supply shocks. The shocks are derived from the structural vector autoregressive (SVAR) model. In a sensitivity analysis, we use unanticipated oil supply shocks derived from oil supply outages as an alternative instrument for fuel prices.