Futron did a great study in (I believe) 2003 that looked at the price elasticity for space launch, and all established markets had essentially none. The reason is that, for military payloads, telecom sats, etc, sticker price of the launch is 3-5% the total system cost (difficult to believe, but true). By comparison, things you could blow at launch up account for 6-10X the cost of the launch itself. Add to that the cost of capital on a multi-billion $$ piece of hardware, and it doesn't take much to realize that the 2 R's - reliability and responsiveness - are much more important than sticker price. The study also found a small but reliable human space flight market, on the order of 5-50 per year sustainably at going costs. It looked at several other options - on-orbit servicing, research, manufacturing - and did not find a business case closing in the near to medium term.
The worst thing - the elasticity is so low that, if you lower your prices the added demand does not keep up. Your revenue goes down as you lower prices! The market actually solves to make space launch as expensive as possible!
It is somewhat bleak, but it also somewhat "IBM broadcasts demand for computers at 2 mainframes per year". I have been able to identify three possible ways out of the trap:
1) Demand elasticity is not the only type of elasticity. There is also supply elasticity, and space has much better characteristics there. Supply elasticity is the inverse of demand elasticity - it is the change in price versus change in units supplied. Launch rate is the most important variable for driving down prices. It's not all paying forward either - with launch rate comes responsiveness, which delivers more value so you can charge more. It also means reliability if you are geared to meet the launch rate (if not you're just doing sloppy work).
2) Reusability. Most people think this only applies if you get a stage back intact, but there is a gradient of reusability. The advantage to spaceflight is probably far greater for recovery of parts. If you can get several engines, turbopumps, computers, and other cost drivers back in good enough shape to do really good post mortems, you will be able to improve both the R's in a remarkable manner. This is far more important than simply trading out manufacturing cost for refurbishing cost early in the industry. Having actual hardware to autopsy against telemetry is extremely useful and would make rocket engineering a little more like automotive engineering and a little less like surgery.
3) Feedback. It is currently best to build extremely expensive satellites, because launch is unresponsive and there is no in-space infrastructure to repair mistakes. The Futron survey did not force execs to consider how they would change their own engineering and development processes in a world where launch was 1/2 or 1/10 the cost - it was mainly asking what the advantage to the business would be for launching current payloads. With high-volume turn-key launch solutions, it starts to make more sense to standardize and modularize payloads. This is good in two ways - first, the payloads are cheaper. That's a good thing simply because when you divide the cost of payload, you multiply the elasticity of the launch by the same factor. If my payload goes from 5% to 20% of total system cost, I will start bargain shopping. Second, it means quantity over quality, and that is great for an industry where launch rate is the primary driver. It leverages demand into supply elasticity.
