But is that actually an argument that is given? Is there data showing that taxis are a significant contributor to congestion?

If this were a real problem, it would be better to estimate the marginal cost to the city for having an extra car on the road for an extra day (or minute or whatever) and charge that. Then, let the market decide how that affects the number of taxis. (Also, if this were true, it would be an argument for charging everyone, not just taxis, to drive in the city.)