Yep. It's Uber that's to blame.

Public transit ridership in major US cities has been flat or declining over the past few years. Several authors have attempted both to explain this trend and to offer policy recommendations for how to respond to it. Past writing on the topic is dominated by theoretical arguments that identify possible explanations, with the few empirical analyses excluding the most recent data, from 2015-2018, where the decline is steepest. This research conducts a longitudinal analysis of  the determinants of public transit ridership in major North American cities for the period 2002-2018, segmenting the analysis by mode to capture differing effects on rail versus bus. Our research finds that standard factors, such changes in service levels, gas price and auto ownership, while important, are insufficient to explain the recent ridership declines. We find that the introduction of bike share in a city is associated with increased light and heavy rail1 ridership, but a 1.8% decrease in bus ridership. Our results also suggest that for each year after Transportation Network Companies (TNCs) enter a market, heavy rail ridership can be expected to decrease by 1.3% and bus ridership can be expected to decrease by 1.7%. This TNC effect builds with each passing year and may be an important driver of recent ridership declines.  

But it's not JUST Uber/Lyft. It's the failure of transit agencies to respond TO the competitive pressures provided by Uber/Lyft. Providing 'circulation' on transit has largely been a myth (Chamber of Commerce dreams of shuttles and loops and circulators are just that). For 95% of people, it's not worth the wait to save the walk).

The idea that it's good for your bus to stop every block is stupid. Every stop costs time, and choice riders choose on time (and reliability). More stops also impair reliability: With so many stops, each stop has (on average) fewer people. But there is no way to predict which stops will have passengers to pick up and drop off at what times. 

To be successful, public transit agencies need to focus on what mass transit can do, that Uber/Lyft cannot: moving a lot of people very quickly. Focusing more on mobility, and less on access.

And how to identify WHERE to cut such services? Easy--just check the map of where the Uber pick-ups are.

source

Where the Uber pickups are, someone has gone wrong in transit provision: Maybe too much is being provided, maybe too little. Maybe it's at the wrong hours, maybe it's going the wrong place. (Uber is expensive, and people don't use it for kicks). 

Uber and Lyft are NOT transit

TNC are  not mass transit; it's a taxi. (There is no massification)

Shared ride _is_ transit. It's a jitney. Jitneys also follow routes. 

The idea that you can combine the 'on-demand' dial-a-ride of a taxi with a jitney is fallacious.

How much detour are you willing to tolerate? And when does it make sense to stop and add another person.

Massification relies not only on spatial convergence, but also temporal convergence: 

Same people in the same place at the same time. 

Which implies it will mostly take place during times of peak demand. 

Which becomes the classic transit agency problems: Sufficient peak capacity is excess off-peak capacity.

The real innovation of TNC's is not the 'dial-a-ride'. It's the ability to use 'surge' pricing to manage the match between supply and demand. It both reduces demand (through higher prices) and draws in additional drivers (through higher fares). 

This is core to the whole TNC model, as it draws in latent supply.

'Latent supply' is how the whole TNC thing works. Most cars are parked most of the time.

Commuting in the morning, the commute+shopping in the evening.

On Ride-sharing

"I think the future is actually in ridesharing. So for example, if you compare the ridership of the Southeast corridor, if you look at the person miles that are traveled on that corridor compared with the person miles traveled on I-25 and I-225 that are right next to it, light rail is carrying about 4 percent. Which means you just need to add one rider to about every 20th car, and you could carry the capacity of light rail. And so when you think about these ridesharing apps and so forth, there's a lot of unused capacity on highways. I think that's where the future is really going to go, is to use the automobile much more effectively and sensibly."

-Michael Ransom- BYU professor 

However. As Mike Brown of Metro Analytics put it:

“BRT would be great, except for the fact that I am only using the BRT corridor for half of my trip.  The rest of the trip involves to/from my house, which is a mile from the corridor, and the other end likewise is 3-4 miles off corridor.”'

Ride-sharing would certainly reduce the last-mile issue. Fewer people per vehicle, fewer detours, less delay per passenger.  But that requires letting a stranger get into your car . Uber and Lyft have proven this can be overcome. But Uber-Pool has not been an enormous success. Which suggests that for most people, the financial benefits of car-pooling aren't worth the time-costs.