Interesting question, but I think part of the answer is fairly obvious, isn't it?

Parasites are often highly specialised to become completely dependent on the host, and in extreme cases even begin to lose basic  body form and structures essential to free-living existence (I'm thinking about cirripedes like Sacculina, but there are many others). Once structures are lost (they have effectively co-opted a foreign organism to act as part of their own body), it becomes extremely difficult to regain them, or equivalent structures. It is also difficult to imagine how it could be an advantage for a successful parasitic organism to begin to regain autonomy. Imagine the changes that Sacculina would need to go through in order to be free-living, and then consider whether evolution would have any cause to select those features in an organism with a parasitic lifestyle.

In contrast, reliable exploitation of a food source is always going to be selected, if it provides a safe environment with high chance of reproductive success; it doesn't surprise me at all that parasitism has evolved so often.

There's probably a huge amount more to consider in this question; I imagine that some lineages have flirted with parasitism, and then gone back to free-living. But it seems fairly obvious (if difficult to prove concisely!) that the more specialised the parasite becomes, the harder it is to regain autonomy.

It would be difficult (if not impossible) to provide convincing evidence that a parasite has evolved to become free-living. It is easier to speculate the other way around.

Interesting question, but I think part of the answer is fairly obvious, isn't it?

Parasites are often highly specialised to become completely dependent on the host, and in extreme cases even begin to lose basic  body form and structures essential to free-living existence (I'm thinking about cirripedes like Sacculina, but there are many others). Once structures are lost (they have effectively co-opted a foreign organism to act as part of their own body), it becomes extremely difficult to regain them, or equivalent structures. It is also difficult to imagine how it could be an advantage for a successful parasitic organism to begin to regain autonomy. Imagine the changes that Sacculina would need to go through in order to be free-living, and then consider whether evolution would have any cause to select those features in an organism with a parasitic lifestyle.

In contrast, reliable exploitation of a food source is always going to be selected, if it provides a safe environment with high chance of reproductive success; it doesn't surprise me at all that parasitism has evolved so often.

There's probably a huge amount more to consider in this question; I imagine that some lineages have flirted with parasitism, and then gone back to free-living. But it seems fairly obvious (if difficult to prove concisely!) that the more specialised the parasite becomes, the harder it is to regain autonomy.

Dear Botting thanks for your great informations  

Miles B Markus misses the point that is key: phylogenetic information provides the basis for asking and answering questions of this sort. Indeed, if there is a lineage that is parasitic through its origin and radiation, but has one (or more) daughter lineage that is free-living, the evidence would be clear. As Joseph Botting points out, there are "borderline" parasitism cases (think about hummingbird-flower mites or hippoboscid flies) that probably flip-flop back and forth. Anyhow, the key is PHYLOGENY! ATP

The genetic characterization studies with which I am familiar or have been involved in have not proven anything in the context concerned, although one can speculate. That aside, I agree that phylogenetic information is (or would be) essential, even though I did not mention this in my short answer. As for hippoboscid flies, some seemed to have happily moved from birds to my head (temporarily) a couple of times. So by extrapolation, I can appreciate that they might be able do some flip-flopping of one sort or another (evolutionarily); starting with their diet, perhaps.

One should first check the situation statistically. The number of parasitic species on earth is at list one order higher than the free living species number. Each free living species carries dozens (or hundreds) of parasitic ones, some of them are specific. Therefore, the rate of free-living species evolving to parasitism has to be (at least) very much higher than the opposite direction. The evolutionary consideration for this situation is clearly explained, up here by J. Booting. The main evolution direction is from generalized forms to more specialized ones. It is like to ask why more bee species evolve from a generalized type to a specialized form then the opposite direction. Under changing environments the specialized species doom to extinct, and then the surviving generalized species evolve to new specialized species.