The argument seems politically motivated. Lower taxes for the consumer tax brackets would be another valid way of increasing GDP (through an increase in consumer spending).
And the graph is meaningless at best. We'd expect to find a correlation b/w the average tax rate across the board and tax revenue as a percentage of GDP. Plotting just the top tax bracket in an effort to show no correlation is intentionally misleading.
And even if he had discovered a valid correlation for this period in this country, there is no reason for calling it a 'law'. There is no indication that the correlation will hold up in the future or in the rest of the world.
And the graph is meaningless at best. We'd expect to find a correlation b/w the average tax rate across the board and tax revenue as a percentage of GDP. Plotting just the top tax bracket in an effort to show no correlation is intentionally misleading.