In this hypothetical scenario, I posit that there will be a significant incentive for large pharmaceutical organizations to pool their resources for research. They do research, even lacking IP law, because they want new treatments to sell. They pool their resources because otherwise their large competitors would use their work without any compensation. This way, they all get access to the research and they all share its costs. Perhaps the ones furthest along take the lead and thus get some form of monopoly rights inside this pool. The agreements made by these entities do not bind, in any way, entities (I posit it will be the small to medium size ones) that are not part of the agreement.
Another thing you overlook is the role of trade secrets. Any entity is free to try to keep any or all information secret. Corporate espionage is still a violation of property rights, and I feel you underestimate the time and effort required to reverse engineer a drug. Even if it's only a matter of months, that's still a huge advantage for the initial developing entity.
The justification for compensation derives from consent, not work. If you dig a hole in my lawn, I owe you nothing and in fact you may owe me damages. If I hire you to dig a hole in my lawn, I owe you whatever amount I agreed to pay you.
They want new treatments to sell, except that they can't sell them, not for profit anyway. So why would they want new treatments? Anyone with enough money to build a factory can just take the research and make the product cheaper than anyone who participated in the research.
Most, if not all, big pharma today already do reverse engineering of the competitors products to learn something that might be useful for their own research. I don't underestimate it I'm afraid. And a few months is peanuts compared to the 20 years or so it takes to find a good drug through research. The time it takes to reverse engineer the drug is about the same time it takes for someone to make it known to the customers, through advertising.
I think you did stumble on the reson for IP in the first place in your comment above. Trade secrets. With no IP I would be a fool to try to capitalize my inventions. Better to keep them secret and make money that way. Let's say I invent a drug that cures every disease known to man, I wouldn't put it on the market, I'd keep it and only give it to those who work for me. That way they'll never be sick and I'd make tons of money more than my competitors. Until someone else figures it out, which may or may not happen. That's the original thought with IP. You get a temporary monopoly for your invention, in exchange for sharing it with the world. It doesn't work flawlessly, but it works.
The thing about IP is the "I" in it. It's hard, if not impossible, to use analogies to physical property. They're just different beasts. But let's say I write a book. You then feel that you have the right to all the hours I spent writing and researching it. Not writing your own book about the same subject, but to take my work away from me. That's not right.