Sleepers are a relatively small part of Amtrak’s ridership. And there’s no guarantee any congress people would even care about room fares. In Amtrak’s view it’s a premium class of service and there is no duty per the mission statement to make sleeper fares cheap and accessible - they feel they have the right to charge what the market will bear unlike the coach fares. There is no guarantee congress people would disagree with that view and demand lower room prices - they have never really weighed in on sleepers vs. the mission statement.
So, as a part of Amtrak's overall ridership, you're right. And of course, this explains Amtrak's occasional/alleged "over-focus" on the NEC (which has historically been around 35-40% of ridership, while state corridors have been around 50% [and on a modest upward trend]) vs the LD trains (which have been about 10-15% of ridership). However, note a chicken-and-egg problem here: The LD trains have
functionally not received a capacity expansion in a
very long time [1], while they've suffered equipment diversions (e.g. Superliners on the Surfliner). while the NEC got the Acelas in the 2000s (and more recently, the Acela IIs have been ordered). Regional trains have also gotten equipment supplements in the form of the California cars/Surfliners and then the Siemens sets.
Within the LD side of things, when we were getting broken-out data, sleeper ridership was generally something like 18% of LD ridership, but it varied a
lot by train. The LSL had a fairly low sleeper ridership share (because of heavy coach turnover in NY), and so did the Star (coach turnover in Richmond, Raleigh, Orlando, and Tampa), some of the western trains were rather higher (as was the Capitol Limited, IIRC), and of course the Auto Train sat at like 40%.
The bigger issue is Amtrak's (apparently voluntary, annoyingly habitual) idling of sleeper equipment and failure to procure more. For clarity, the CAF order "went bad" for reasons not entirely within Amtrak's control (there was a "low bidder"/"inconvenience discount" at issue), but in general Amtrak hasn't seen this as a priority. The split fleet problem has also been a devil here, since inevitably any order is only for half of the trains (this is why I'm a fan of going to an all single-level fleet - I'd like to be able to swing 30-50 sleeping cars off of the Western Transcons in winter
and put them on the Florida trains rather than just losing the capacity [2]).
As a note, sleeper ridership was something like 600-700k/yr for a long time. I don't know what it is now, but Amtrak's capacity constraints don't help there. I'd guess it's closer to 500k now (give the number of trains that have lost a sleeper). And yes, I do think that if the capacity were available, yield factors might be a bit lower but sleeper ridership would probably land somewhere in the 800k-1m range.
There's one reason to care about sleeper access, and that's access to the dining car. That is at best erratic for coach pax now (on a number of routes it is non-existent, and on others access is touchy-at-best), which gets problematic on longer trips. Now, the solution
there is probably to order Amtrak to make diner access available to coach pax, but that may require some percussive therapy.
One way of getting a feel for what you're getting for your $ is to calculate the cost of travel per hour of travel. F'rinstance, I just picked the random date of 17 April to compare a Roomette with flying for two people (each with one checked bag) from CHI to SEA.
• EB worked out to $23.66 per hour
• AA or Delta non-stop worked out to $43.36 per hour.
Even a high bucket Roomette for two beats flying using this metric at $40.22 per hour. So if speed's not relevant, . . . . . .
That's one
hell of a metric to go by given that the average traveler is, at a minimum, not likely to consider spending lots of time
en route a positive (even if they don't consider it to be a negative). Also, is that a coach traveler on those flights, or a passenger in low-bucket first class? I strongly suspect the former, and I think a lot of us have generally objected to using an airline coach fare vs a sleeper fare.
[1] The Viewliner II order has not acted to increase capacity, though the blame for this is with Amtrak idling a lot of equipment. The Viewliner I order was actually a decrease, since it was replacing pre-Amtrak equipment at less than 1:1. So the last expansion of capacity was either from the Superliner IIs (and it was a temporary increase), the Superliner Is (there's a case this was temporary since it allowed a lot of older stuff to be phased out), or...never (the Amfleet IIs were also Heritage replacement). I think the Superliner Is
do count, FWIW, but the increase was marginal.
[2] I did an analysis of monthly sleeper ridership data (I posted my findings in a thread here some years ago). Essentially, the Western trains are
very seasonal while the Florida trains plus the Crescent tended to have near-flat ridership (suggesting that any seasonal variation in demand was simply getting blocked by a brick wall of capacity constraints).
