A private taxi causes as much congestion as an occupied car on the same trip, plus wasted congestion whenever it is empty (no passenger).

It can only reduce congestion if multiple passengers on different trips are in the car at the same time.

> It can only reduce congestion if multiple passengers on different trips are in the car at the same time.

Not _only_: in the general case, there's nothing stopping taxi services from providing a last-mile (or rather last-n-miles) connection to transit compared to a car trip for the entire time. Depending on the way this works out, congestion can be reduced with private taxi rides. In fact, Beverly Hills city council passed a resolution to investigate driverless shuttles for exactly this purpose and precisely to reduce congestion.

Although 1 person driving another person from point A to point B reduces the number of single-occupancy vehicle trips, it doesn't necessarily reduce congestion. It could even increase congestion because all of those "private shuttles" are putting in so many miles on the roads when people would otherwise take higher-occupancy options or drive themselves (fewer miles, but that dreaded single-occupancy vehicle trip count goes up).

Well you're in luck, because that's why traffic engineers measure vehicle-miles-traveled and private shuttles saved almost 29 million miles.

On a personal note, I don't really understand what you're saying. Congestion goes up because there are less vehicles on the road? It sounds like you're trying to play contrarian and arguing for unintended effects and trying to assert something about short trips, but I'm at a loss for how your hypothesis is supposed to work.