In attempting to answer this, I read two different questions here.
- Why is it so hard to develop a space vehicle/rocket?
- In contrast, why did SpaceX and Oribtal look so easy?
Thus for #1, because it is rocket science. For #2 There is one word to answer it. Heritage. The best example in modern history of how that heritage manifests is NASA. The second example of heritage that is less specific is the military.
US corporations can publicly leverage all the research NASA did in the 50s till now.
When SpaceX designed the capsule shape for Dragon, they could look at all the work NASA did in the Mercury/Gemini/Apollo days on different shapes, see the results and make a better decision. Much less 'testing' on their own that they needed to do.
For parachute design for Dragon, same thing. Apollo did a TON of parachute testing, and SpaceX leveraged all of it.
Their engine designer, Tom Mueller, came from TRW where he worked on rocket engines there. He was able to reuse the basic design approach of the Lunar Lander ascent engine for a simple engine design, then cluster it for performance.
To be clear, SpaceX has said this loud and clear on many occasions that they stand on the shoulders of giants and would not be possible without the research NASA had done over the decades. (I.e. The actual purpose of NASA!)
Orbital cheated, and used the overseas heritage equivalents of NASA. They outsourced it. They hired the guys who build Zenit (Yuzhnoye) to build their first stage. (Zenit is also LOX/Kerosene powered, and uses an RD-170 (an RD-180 is 'half' of an RD-170, and the RD-191, used in Angara, is 1/4 of an RD-170)). They cheated further by reusing engines from the Soviet era N-1 program, that were already built, and in storage. Then, they used existing Castor engines from ATK. (I use cheat in a non-perjorative fashion here. Buying off the shelf is just smart business. I question the use of the NK-33/AJ-26 engine, but still).
Delta 4 uses a newly developed core, newly developed engine, and that was hard and not cheap. But that was Boeing, who had acquired companies with all the expertise and IP in house already. (That may be a LH2 issue, liquid Hydrogen is just plain hard to work with).
Look how much trouble India and China had developing indigenous launchers, when they could not piggy back on NASA or Russian experience. Conversely, look at Shenzou which is a major clone of Soyuz.
Israel with the Shavit did what the US/Soviets did in the 1950/1960s (Atlas, Delta, Thor) and re-purposed an ICBM as a space launcher. (The Russian Soyuz booster (as opposed to Soyuz spacecraft) is basically the R-7 ICBM from the 1950's with upgrades).
ESA did not have any real ICBMs to tag off, and you see that in Ariane, being more of a fresh design. Which was not easy either.