The fuel for a Falcon 9 (SpaceX) costs around \$200,000 per launch, while the launch itself costs \$62,000,000. Safety precautions, rocket shell and rocket engines cost a lot. But why not fuel?
As noted by others, Methane, RP-1, LOX are manufactured in large scale across the US and much else of the world. (Manufactured may be the wrong word, perhaps produced is better).
It is functionally a commodity.
On the other hand, aerospace engineers get relatively high salaries. Space rated components, come from a world where you get one shot on a launch and it better work absolutely perfect. That sort of reliability is much more expensive, than simply processing LOX by compressing and cooling it.
SpaceX employs several thousand employees. The launches generate most of the money to fund their salaries. The reason the Space Shuttle cost so much was not the hardware per se but rather the 20+ thousand employees directly considered working to ensure each launch.
Lean implies smaller workforces.
Interesting anecdote to which I cannot find a reference. While a Falcon 9 launch lists at about \$62 million US, a launch for the Air Force costs almost \$100 million. Why? Because of additional paperwork that the Air Force requires. I.e. They spend almost \$40 million on paperwork! That is an amazing amount of paper.
Of course it is the cost of employee time to fill out, and process the paper, and the additional tasks it implies. Which should terrify you at how much work it must be to generate \$40 million worth of paperwork.
Making fuels is a matter of relatively cheap bulk liquid processing from relatively cheap feed-stock. Yes RP-1 is more expensive than JET-A because more processing is needed and it's more specialized but it is still bulk liquid processing.
LOX is even cheaper because the feedstock is atmospheric air, there are just the processing costs.
Making aerospace components involves expensive processing of exotic materials. This processing often happens in low volume so there are relatively few economies of scale.
For comparison, 747 apparently costs somewhere around \$350,000,000 while a full fuel load for one is more like \$100,000. So the cost of a full tank of fuel being much smaller than the cost of the vehicle is hardly unique to space rockets.
The difference is that a 747 can be used tens of thousands of times. So over its lifetime the fuel costs more than the plane. Rockets on the other hand are typically only used once and when they can be reused refurbishment costs are a concern.
What is cheaper: buying a new car every time or refilling the tank? Also it depends on the fuel: oxygen is pretty cheap compared to hydrogen (about 20 times more expensive). This is one of the reason that BFR is going to use methane and oxygen as fuel- they are both very cheap (also the abundance of carbon dioxide which is straightforward to convert to methane on Mars).
A point that other answers have not covered: there aren't any better-yet-more-expensive fuels that don't have huge disadvantages. A chemical rocket engine is a fairly simple machine in principle (the details can get complicated) and the things you most want in a fuel/oxidiser combination for space launches are:
- Lots of energy released in the combustion.
- Low molecular weight of the combustion products.
Liquid hydrogen and liquid oxygen have the best performance in theory, but liquid hydrogen has very low density, making the tanks for it heavier. RP-1 or liquid methane give you much better tank weights, and you end up with a more effective rocket. Liquid fluorine might give you more energy than oxygen, but the safety problems would be horrible.