Market forces encourage restraint. An attorney has an incentive to help smaller clients spend money only where it will help them make their business successful. Once the business is successful, they will become a bigger client, and the attorney will hopefully get more work from them because they treated them fairly when they were small. A rule of thumb is that it's much easier to bring in more business from existing clients than it is to go out and find brand new clients.
You could also think of it like the fishing industry: A fishing captain wants to bring in as many fish as he can, but if he takes too many, he runs the risk of underpopulating the fishery and destroying his livelihood. The fishing captain, like the attorney, has an incentive to ensure that the source of his income remains healthy.
Also, at least in Seattle, the patent bar is fairly small. An attorney that goes around convincing people to spend a lot of money filing patent applications that go nowhere will probably earn a reputation for doing so, and would have a harder time acquiring new business.
At least one case shows it may be hard to win a malpractice claim in court. It seems like succeeding on a claim alleging that an attorney advised their client to pursue a patent on something unpatentable would be difficult.
I did a little searching on this, and I found Voight v. Kraft, 342 F. Supp. 821 (D. Idaho 1972). You can read the case here:
Summary: Plaintiffs tried to sue their patent attorney in federal court for advising them to pursue a patent on an unpatentable invention. However, the federal courts only have subject matter jurisdiction if the case contains questions of patent law. As the plaintiffs admitted that there was no question of patentability, there was no federal question, and so the case was dismissed for lack of jurisdiction.
That was the only case I was able to find in a brief search on Google. However, I did find this article by Professor Oddi that goes into great detail about the difficulties involved in proving patent malpractice:
From the patent attorney's perspective, advising that an invention is not patentable is often a bad idea. An attorney can definitely be sued for malpractice for negligently advising a client not to pursue IP protection (or negligently failing to obtain such IP protection), and the client suffers damages.
If you assume that the attorney will get sued no matter what advice he gives, think about this: What would the damages be for failing to obtain IP protection that could have ensured a stake in a billion dollar industry? Meanwhile, what would the damages be if a client spent legal fees on patent prosecution to pursue an application that is unlikely to be allowed? If the attorney is erring on the side of his own risk management, it's probably in his best interest to advise the client to try to seek IP protection.
So, in answer to the question, it seems like the risk of malpractice liability provides an incentive to advise pursuing patent protection in questionable cases.
Personally, I don't like to tell clients that they shouldn't file a patent on something, because I can never be sure whether it is patentable, and in the end it is their decision to make. However, in difficult cases, I like to manage expectations by explaining how hard/expensive I think it will be to obtain a patent, and predicting that any claims eventually allowed are likely to be quite narrow or of little value.