With only one month left in the fiscal year, I reran my projections which attempt to back out the overhead (based on the latest data we have on the overhead, which is years old) to get the profit/loss based on *direct costs*. (To do this a month early, I used variances between last year and this and just assumed September would have the same numbers as last year, which is as good as I can do until the September report comes out.) To be clear about this, if a train is profitable based on *direct costs*, then cancelling the service would mean Amtrak loses more money every year. These trains contribute to Amtrak's bottom line. They just don't contribute enough to cover very much overhead.
My estimates do not account for any possible overhead reallocation which occured due to the Silver Star having its dining car removed. If it had its overhead costs reduced, then that means that *the Silver Star is actually doing significantly worse than my estimates*. If that overhead was allocated elsewhere, then I*other trains are actually doing better than my estimates*.
Furthermore, if Amtrak has massively reduced overhead in the last few years (which would be a good thing), then all the trains are individually doing worse than I am estimating. If Amtrak has substantially increased overhead, on the other hand (...which is more likely due to inflation), all the trains are doing better than I think.
My current estimates, and they remain estimates, are that:
-- The Auto Train is generating about $38 million / year to help cover overhead
-- The Silver Star and Silver Meteor are each generating about $15 million/year
-- The Palmetto is generating about $13 million/year
-- The Crescent seems to have gone into the black, generating about $3 or $4 million/year. (Interestingly, this matches with a projection I made last year for when it would become profitable based on simple averaging of trends since 2012.)
-- The Lake Shore Limited is still generating about $2 or $3 million/year
-- The Coast Starlight seems to have gone into the black, generating about $1 or $2 million/year. (Interestingly, this matches with a projection I made last year for when it would become profitable based on simple averaging of trends since 2012.)
-- The Empire Builder is probably fluctuating a bit short of breakeven. My raw model says $1-2 million in the red, but I know its costs increased because BNSF is getting its OTP payments again which it wasn't before, so it's probably $3 to $6 million in the red. (My projection based on trends says it should be in the black in in 2017.)
-- The Capitol Limited is still in the red by $4 to 5 million. I believe the Cap generates substantial income on other trains through connecting traffic, so in fact it probably would cause Amtrak to lose more money if it were cancelled.
-- The CONO is still in the red by $3 to $4 million
-- The Cardinal is still in the red by $3 to 4 million. (Same projection as used for Crescent and Coast Starlight says it's in the black in 2018, updated says 2020, but if it were daily it would be profitable today.)
-- The SWC is still in the red by $6 to $7 million. (Same projection as used for Crescent and Coast Starlight says it would be in the black this year, but I was wrong. Updating the projection gives 2020.)
-- The Texas Eagle is still in the red by $6 to $7 million. (Same projection as used for Crescent and Coast Starlight says it's in the black in 2019, updated projection says 2021.)
-- The CZ is in the red by $8 to $9 million. (Same projection as used for Crescent and Coast Starlight says it's in the black in 2018.)
-- The Sunset is in the red by $11 to $12 million. (Same projection as used for Crescent and Coast Starlight says it's in the black in 2021, but if daily, by 2019.)
The key takeaway from this is how much better every one of these trains is doing financially than they were in 2012.
Other conclusions I take from this:
-- As I've said before, nearly all of Amtrak's operating subsidy goes to overhead. More trains == more gross profit to cover the overhead.
-- The Viewliner dining and sleeping cars, *if they ever deploy*, will mostly cut the running costs and add to the revenue of trains which are already profitable before overhead.
-- The entirety of the "subsidy for running particular trains" on the long-distance network (as opposed to "subsidy needed to have a train system at all") goes to the four Western Transcons, the Texas Eagle, the CONO, the Cardinal, and the Capitol Limited (which is almost certainly generating enough revenue on connecting trains to cover its subsidy).
-- The EB, CZ, and SWC should be profitable before overhead around 2018 just based on trends.
-- The Cardinal would be profitable before overhead right now -- about $4 to $5 million in the black -- if it ran daily. (This assumes costs multiply by 1.5 and revenues multiply by 7/3.) If CSX decided to charge an extra $4 million/year for the "privilege" of running daily, it would *still* be worth it to run daily. The increased profits of $5 - 6 million / year are certainly enough to make a rational pitch for an RRIF loan to build $30 million in improvements if that's what it takes to run daily.
-- A daily Sunset Limited should be profitable before overhead around 2019.
-- This would mean the only long-distance trains with subsidies for direct costs would be the Texas Eagle, CONO, and Capitol Limited -- and the Capitol Limited is generating connecting traffic sufficient to cover iits subsidy.
At this point the political pitch to Congress has to be "Basically all the trains make money, just not enough to cover overhead. Cutting service will cost you more money; adding service will cost you less money. We're just asking you to cover our overhead costs." I think this pitch should be hammered hard, because it makes it very hard for them to yammer about "money-losing trains". Congress is not subsidizing the operation of the trains. They're subsidizing the ticketing system and the cost of mainting stations and the central dispatching offices and the cost of catching up on deferred maintenance by the former private railroads. They're subsidizing the equivalent of the airports, the air traffic control system, the road maintenance crews, the highway patrol.
