Per-mile pricing is the native language of trucking. A carrier quotes $2.60 a mile, you multiply by the distance, and you have a number you can put in a budget. Shippers moving into rail reach for the same handle, and it mostly doesn't fit. Rail does not price by the mile, it prices by the car; the mile figure is something you derive afterward. Derive it wrong — wrong denominator, wrong distance, wrong scope of costs — and you will either talk yourself out of a lane that works or into one that doesn't. Here is how to build the number correctly and what it is actually good for.
Three Different "Per Mile" Numbers
When someone says rail freight costs a certain amount per mile, ask what they divided by. There are three answers in common use and they are not interchangeable.
| Metric | What it measures | Use it for |
|---|---|---|
| Cost per car-mile | All-in cost for one railcar, divided by route miles | Comparing lanes you actually run; sanity-checking a quote |
| Cost per ton-mile | Cost per car-mile divided by net tons loaded | Comparing rail against truck or barge; talking to finance |
| Cost per loaded mile | Total round-trip cost divided by the miles the car ran loaded | Private and leased fleets where you pay for the empty return |
A single grain move can be $4.00 per car-mile, 4 cents per ton-mile, and — if the car comes back empty on your dime — closer to 8 cents per loaded ton-mile. Same shipment, three defensible numbers, three very different stories. Most bad rail-versus-truck analysis we see is not a math error. It is one side of the comparison quoted in ton-miles and the other in vehicle-miles.
Ton-mile is the industry's default for cross-mode comparison, and it is the one that shows up in published statistics. Car-mile is the one that matters operationally, because the car is what you are actually buying — a point covered in more depth in how to estimate rail shipping rates.
How to Calculate Rail Cost Per Mile
Three steps. Do them in this order, because each one depends on the last.
- Build the all-in cost for one car. Line-haul rate, plus fuel surcharge, plus every short-line segment on the route, plus switching, plus car lease or per diem if you supply the equipment. Not just the Class I line haul.
- Divide by the rail routing miles. That gives cost per car-mile. Use the miles the car will actually travel over the network, not the distance between the two cities on a map.
- Divide by net tons loaded. That gives cost per ton-mile. Use what you actually load, not the car's rated capacity — if you load a hopper 12% light, your real cost per ton-mile is 12% higher than the brochure math.
A worked example, using illustrative figures rather than a quote for any particular lane:
- Covered hopper of grain, all-in $4,800 per car, 1,200 rail miles, 100 net tons loaded.
- Cost per car-mile: $4,800 ÷ 1,200 = $4.00.
- Cost per ton-mile: $4.00 ÷ 100 = 4.0 cents.
Now the same commodity on a much shorter lane. Say $2,400 all-in over 300 miles, same 100 tons:
- Cost per car-mile: $2,400 ÷ 300 = $8.00.
- Cost per ton-mile: 8.0 cents.
The short lane is half the total dollars and double the cost per mile. Nothing went wrong — that is simply what the rail cost curve looks like, and if you carry a single per-mile figure across lanes of different lengths, you will misprice every one of them.
Which Distance? Rail Miles vs Highway Miles
Use the rail routing distance. This is the step people skip, and it quietly inflates every per-mile number they produce.
Rail routings are rarely the shortest line between two points. The network moves cars through junctions, classification yards, and interchange points that exist because of how the railroads were built and merged, not because of where your freight is going. A car can travel materially farther than a truck covering the same origin and destination — and on a fragmented route with two or three carriers, farther still, because each carrier hands off where its own network allows. Our guide to railroad interchange covers why those handoff points sit where they do.
The practical failure looks like this. A shipper prices a move at 900 highway miles, gets an all-in car cost, divides, and reports a per-mile figure. But the car is routed over more than 900 miles of railroad. The denominator was too small, so the per-mile number came out too high, and rail looked worse than it is. If you are benchmarking one lane against another, this error is not symmetric — it punishes circuitous routings twice, once in the real cost and once in the arithmetic.
When you don't have routing miles in hand, say which distance you used. A per-mile figure labeled "highway miles" is a usable approximation. An unlabeled one is a trap for whoever reads your spreadsheet next quarter.
Why the Per-Mile Number Falls as the Haul Gets Longer
Rail cost per mile drops steeply with distance because most of what you are paying for happens at the ends of the move, not in the middle.
Placing and pulling the car, switching it in and out of a yard, classifying it into an outbound train, handing it to a connecting carrier, and spotting it at destination — that work is essentially the same whether the car then travels 200 miles or 2,000. Line-haul movement, by contrast, is where rail is extraordinarily efficient: a single locomotive consist moving a hundred loaded cars is the cheapest ground transportation in North America on a per-ton basis. Stretch the haul and you dilute a fixed block of terminal cost across more of the cheap part.
Illustrative shape of that curve, holding commodity and car type constant:
| Lane length | Relative cost per car-mile | What's driving it |
|---|---|---|
| Under 200 miles | Highest — often at or near a rate floor | Terminal and switching cost dominates; little line-haul to amortize it |
| 200–600 miles | Falling fast | Line-haul share growing; truck still very competitive here |
| 600–1,200 miles | Flattening out | Rail's structural advantage on bulk starts to show clearly |
| Over 1,200 miles | Lowest and fairly stable | Terminal cost is a small fraction of the total |
Two things follow from the shape. First, a per-mile figure borrowed from a long lane will badly understate a short one — that is the arithmetic behind the effective minimums built into rail rate structures. Second, a lane that fails a rail evaluation at 400 miles can pass comfortably at 1,400 with the same commodity and the same volume. Distance is not a multiplier on rail; it is a qualifier.
Comparing Rail and Truck Without Fooling Yourself
Compare per ton-mile, on the same scope of cost, or don't compare at all.
A truckload rate of $2.50 a mile sounds cheaper than $4.00 a car-mile until you remember that one bulk railcar carries roughly three to four truckloads of the same product. Normalize both to tons and the picture inverts. A dry van or hopper-bottom trailer at $2.50 a mile carrying about 22 net tons works out near 11 cents per ton-mile. The 1,200-mile rail move above came in at 4 cents. Rail is meaningfully cheaper on that lane — but it is roughly a 2 to 3x advantage, not the 5x or 6x you get by comparing against a national trucking average.
Then add back what a per-mile figure leaves out on both sides:
- Drayage and transload. If either end of the move isn't rail-served, you have a truck leg plus handling, and those costs are per ton, not per mile of the rail haul. On a short rail lane they can exceed the line-haul entirely. Our post on transloading vs direct rail works through when that math still clears.
- Working capital. Rail transit runs significantly longer than truck — weeks rather than days on a long move, with variability at every handoff. That inventory is financed by someone, and it belongs in the comparison even though it never appears on a rate sheet.
- Demurrage and detention exposure. A per-mile number assumes the car moves. Cars that sit accrue charges that no per-mile figure predicts. Budget a realistic allowance, and see how demurrage actually works before you assume it's a rounding error.
- Equipment. If you lease or own the cars, lease cost and empty repositioning are yours. That is what turns cost per car-mile into cost per loaded mile, and on an unbalanced lane the difference is close to double.
Five Things That Break a Per-Mile Comparison
In rough order of how often we see them do real damage:
- Mixing denominators. Truck quoted per vehicle-mile, rail quoted per ton-mile, conclusion drawn from the two side by side. This one error has killed more viable rail lanes than any pricing factor.
- Line-haul only on one side. A rail per-mile figure without fuel surcharge, short-line segments, and switching, compared against a truck rate that already includes fuel and door-to-door service. Scope both to the same fence line.
- Ignoring the empty. Shipper-supplied cars on a one-directional lane run back empty. If you pay for that mileage or that lease time, your true cost per loaded mile is far above your cost per car-mile.
- Rated capacity instead of actual loading. Dividing by what the car could hold rather than what you put in it. Light-loading is invisible in a per-car rate and brutal in a per-ton-mile one — which is why picking the right car type is a cost decision, not a logistics detail.
- Averaging across lane lengths. A blended per-mile figure from a network of 300-mile and 1,500-mile moves describes none of them. Build per-mile figures per lane, then aggregate for reporting if you must.
What a Per-Mile Number Is Actually Good For
Used carefully, a per-mile figure is a screening and benchmarking tool — not a pricing tool.
Good uses:
- Screening lanes before you invest analysis time. Apply a rough per-ton-mile band to a candidate lane and see whether it is anywhere near your truck cost. If it's off by an order of magnitude, you learned that in thirty seconds.
- Benchmarking your own lanes year over year. Same lane, same commodity, same car type — the per-car-mile trend tells you whether your pricing is drifting relative to what changed in fuel and volume.
- Detecting a bad quote. If one lane in your network prices at triple the per-car-mile of a comparable one, something specific explains it — a short-line segment, a captive origin, a car-type premium, a circuitous routing. It's a question worth asking, and often worth money.
- Talking to finance. Cost per ton-mile is the unit that translates across modes and into a P&L. Convert at the end, present in ton-miles, keep the operating math in cars.
Bad uses: quoting a customer, budgeting a new lane off a borrowed figure, or estimating a short haul from a long-haul rate. Rail rate structures do not extrapolate linearly, and per-mile figures pulled from someone else's network carry their carrier mix, their commodity, and their volume — none of which are yours.
If you want to go deeper on where rail pricing comes from in the first place, the railroad pricing module in our free rail logistics course covers tariff structure, contract rates, and surcharge mechanics in order. To put real numbers behind your own lane, run it through the rail rate estimator for an indicative all-in figure, or send us the origin, destination, commodity, car type, and annual volume — we will map the carriers on the route, tell you the routing miles, and give you the per-mile math with the assumptions visible. That's the core of what our rail logistics services do.
Frequently Asked Questions
What is the average freight cost per mile by rail?
Rail line-haul commonly works out to roughly 2 to 4 cents per ton-mile once you divide an all-in carload rate by tons moved and miles traveled. Expressed per car-mile instead, the same move can land anywhere from a few dollars to well over ten dollars per mile, because a car-mile figure carries the whole car regardless of how much is loaded in it. Always state which denominator you are using before quoting a per-mile number.
How do I calculate rail freight cost per mile?
Take the all-in cost for one car on the lane, including fuel surcharge and any short-line segments, then divide by the rail routing distance to get cost per car-mile. Divide that result by the net tons actually loaded to get cost per ton-mile. Use the rail routing miles, not the highway distance between the two cities, or your per-mile figure will read artificially high.
Why is rail cost per mile higher on short lanes?
Because a large share of the cost of any carload move is incurred at the ends, not in the middle. Switching, yard time, terminal handling, and car placement happen whether the car travels 150 miles or 1,500. On a short lane that fixed block is spread over very few miles, so the per-mile figure spikes. This is also why rail rate structures carry effective minimums per move.
Is rail cheaper per mile than truck?
Per ton-mile, rail is usually cheaper on bulk volume over long distances, often by a meaningful multiple. Per vehicle-mile the comparison is meaningless, because one railcar carries roughly three to four truckloads of bulk product. The honest comparison is landed cost door to door for the same tonnage, including drayage on both ends and the working capital tied up in a transit cycle measured in weeks rather than days.
Does the fuel surcharge belong in a cost per mile calculation?
Yes, as long as you say so. A per-mile number built on line-haul alone is not comparable to a truck rate that already bundles fuel. Use a 12-month average of the applicable surcharge program rather than the current month, since the programs reset monthly against a fuel index and a single data point will not reproduce.
What is the cost of transporting goods by train versus truck over 1,000 miles?
On bulk commodities at carload volume, rail typically wins on a landed basis at that distance, and the advantage widens as the haul gets longer. The result flips when the freight is small, time-sensitive, or needs door-to-door delivery in a place the rail network does not reach, because the truck legs on each end can consume the entire line-haul saving. Run the specific lane rather than relying on a national average.