There are good arguments for directly subsidizing public transit. Here I analyze whether directly subsidizing public transit is a good thing to do for distributive reasons.
The best way to go about this is with an example. Suppose we have a public transit system that is entirely financed by fares. Someone proposes that we drop the fares by $1/ride by using $1 million of tax revenues to directly subsidize the transit system. Here the example involves increasing subsidies to decrease fares, but all of the conclusions that spring from it equally apply to examples involving decreasing subsidies to increase fares. It’s two sides of the same coin.
The distributive argument for decreasing fares in this way is two-fold. First, transit fares are regressive. One dollar to a poor person is more significant than $1 to a rich person. So by dropping the fare $1, we decrease the amount poor people pay (measured as a share of their overall income) much more than we decrease the amount rich people pay (again, measured as a share of their overall income). Second, poorer people ride public transit more than rich people. So not only does the fare drop funnel more money to poorer people as individuals, it also funnels more money to poorer people as a group because they ride transit more frequently. Thus, this argument concludes directly subsidizing public transit leads to a more egalitarian distribution.
The issue with this argument is that it does not take into consideration the opportunity costs of using that $1 million to subsidize fares. That is, the argument does not consider what else we might do with that money. Consider, for instance, this alternative proposal: take the $1 million and hand it out to lower-income people that ride the transit. This alternative proposal creates an even more egalitarian distribution than the subsidy does. The gains of the subsidy are at least partly captured by higher-income riders because they also enjoy lower fares from it. Through cancelling the subsidy and giving the money directly to poor riders, we ensure all of the gains of the $1 million go exclusively to the lower-income riders. This would improve their plight more than the subsidy would.
Again, I am not saying there are not good reasons to subsidize transit. I think there are, especially when we are talking about the environment and traffic. However, I am generally wary of any argument that says subsidizing a universal service is a good thing to do for distributive reasons. In almost all cases, a better distributive outcome will be achieved through canceling the subsidy and using the savings to target money or other benefits directly at lower-income people. Why? Because a universal service, by its very name, is utilized by all people, and not all people are poor.
Of course, we must also consider the politics of this all. As a society, we seem to be far more tolerant to distributive policies that operate somewhat covertly. If you can somehow bake distributive changes into the cake (see predistribution), people seem not to regard it as a distributive change at all. This is probably why we have the submerged state and why people love minimum wage increases, but seem to be more wary of overt transfers. Given that political background, it may just not be feasible to do the most distributively beneficial thing all of the time. So it may be wise to just take what you can get even if what you can get inefficiently leaks money to the rich, as direct subsidies of universal services do.
But when we make concessions for political feasibility reasons, we should at least understand that we are doing that. We should not get so wrapped up in it that we forget that as a matter of fact, spending a marginal dollar on fare subsidies is not the best way to alter the distribution, just the best way we can manage given the politics of our moment.