The most recent overall study of this that I know of is File-Sharing and Copyright, by Felix Oberholzer-Gee and Koleman Strumpf.

The advent of file sharing has considerably weakened effective copyright protection. Today, more than 60% of internet traffic consists of consumers sharing music, movies, books and games. Yet, despite the popularity of the new technology, file sharing has not undermined the incentives of authors to produce new works. We argue that the effect of file sharing has been muted for three reasons. (1) The cannibalization of sales that is due to file sharing is more modest than many observers assume. Empirical work suggests that in music, no more than 20% of the recent decline in sales is due to sharing. (2) File sharing increases the demand for complements to protected works, raising, for instance, the demand for concerts and concert prices. The sale of more expensive complements has added to artists' incomes. (3) In many creative industries, monetary incentives play a reduced role in motivating authors to remain creative. Data on the supply of new works are consistent with the argument that file sharing did not discourage authors and publishers. Since the advent of file sharing, the production of music, books, and movies has increased sharply.

For my and Ahto Apajalahti's book Jokapiraatinoikeus ("every pirate's right", a reference to the Nordic legal doctrine of everyman's right) I also a collected a number of statistics showing no decline in the production of works, and for many fields, no or only a moderate decline in sales, as well. That suggests that the overall utility gain of illict file-sharing is quite considerable. Unfortunately our book is only available in Finnish, but Against Intellectual Monopoly, which we cited frequently, makes many similar points and covers a lot of the same ground as we did.

You should generally not pirate, because, to the extent that you (and thus similar beings) would regard it as optimal to pirate, potential creators would not regard it as optimal to locate narrow targets in designspace (whose value comes mostly from its incorporeal component), including those cases where the value of locating these targets exceeds the value forgone in finding them (e.g. labor and other resources expended to find the target).

This conclusion holds even and especially for the case where high-value targets have already been located and became public knowledge (via piracy), and thus your current decision of whether to pirate (to assimilate the target's location) no longer has causal influence on whether the target will be discovered.

The reasoning behind the above conclusion is isomorphic to that of why one should "one-box" in Newcomblike problems, where the decision of whether to provide the (additional) utility has already been made, and any individual's decision to refrain from utility extraction cannot causally influence the decision to provide the utility.

To highly compress and simplify the above reasoning into a human-appealing slogan: "Too much piracy means too little Microsoft Word."

Note: the above reasoning does not mean that all measures to stop piracy, nor any particular one, are optimal, for much the same reason that the general optimality of punishing defectors does not imply the superiority of all possible punishments over non-punishment.

Adobe thinks this is stealing because Adobe doesn't think that its marginal cost of producing another copy of the software is low. Sure, the out-of-pocket cost for copying some bits and bytes is low, but Adobe wants to include some amount for the cost of writing the software in its accounting of the cost of providing another copy.

And Adobe is right (and rational) to want to do this. If you can't figure out how to amortize the costs of the creating the product in the first place into the marginal price of another copy, you literally will never think that large upfront costs are a good idea. In other words, we aren't deciding whether the legal regime allows non-payers to get Adobe after the payers get it. We are deciding whether to have a legal regime that creates incentives for the existence of Adobe at all. And that ignores the signalling costs Adobe has in figuring out when it has sold "all" the copies it can.

Obviously, there is an empirical question about how much protection provides the appropriate incentive for the creation of new stuff. But I think there is a far amount of evidence that zero protection does not create enough incentive.

This would be a great topic for Utopia, Dystopia, Weirdtopia

I'll do two versions of the utopia and dystopia.

Utopia 1: Copyright is abolished. Music is produced for a combination of self expression, fame, concert ticket sales, and merchandise sales (merchandise is also largely copyable, but many people buy it from the artist anyway). Movies are mostly crowd-funded through places like kickstarter, some low-budget ones are self-funded initially and then rewarded with donations. Widely used software is mostly open source. Where GPLed software (no longer protected by copyright) is modified and distributed without source by a large company for profit, righteous hackers reverse engineer the program and bad mouth the company publicly, so it's generally seen as not worthwhile. Small GPL violations by individuals are usually ignored.

Dystopia 1: Fair use is abolished, so derivative works can only be produced by the large companies who own the original copyright, or those with permission. Some companies have large copyright portfolios and have an uneasy truce against mutually destructive lawsuits for parodies and occasional references to each other. Positive reviews are often given permission to quote from a work, but negative ones never are. Singing with others present, or playing a CD through speakers instead of headphones, is considered a public performance and requires a license. Letting someone use your e-reader to read your books is forbidden. Sometimes companies sell widely referenced works to copyright trolls who don't publish anything and can freely sue all the other companies who have derivative works without being sued themselves. Anonymous communication on the Internet is banned and putting up an individual website is very expensive, as hosting. General purpose PC's require an expensive permit and monitoring software. iPads and Android devices can only run approved software, and it's much more expensive due to lack of competition and the approval process. Old public domain texts end up effectively owned, as slightly modified versions come into common use, and many key phrases become trademarked after use in advertisements. It goes without saying that Disney's copyrights never expire.

Utopia 2: Copyrights and patents are easy to register and enforce, and micropayments make them easy and convenient to pay for, so most individuals have some registered ideas, jokes, or stories making them at least a little money, without undue burden on the users. Patents have reasonable legal limits on how expensive they can be. Copyrights have voluntary limits on how expensive they can be - overly expensive stuff is ignored and avoided, e.g., many radio stations have a guarantee that any song you hear on their station can be purchased for a reasonable price.

Dystopia 2: the Internet made copyright unenforcable and very few movies and very little music is produced. Most of it is crap. Some is a mixture of good stuff with a just-bearable amount of ads for generic Viagra and fraudulent money-making schemes, somehow made integral to the story. Very little consumer software is produced, though malware is thriving due to common software not getting security updates. Individuals cannot make money from software except through fraud.

ETA - oops, that Utopia 2 may be a weirdtopia for me. The problem is I was trying to come up with a utopia and dystopia from a pro and anti copyright viewpoint, and also Utopia 2 was largely inspired by an sf book I vaguely remember. Possibly by John C Wright.

In the case you have chosen to illegally download a copy of the software, some people would compare that to stealing (certainly the folks at Adobe).

FWIW, I've heard that Adobe doesn't really crack down on illegal copies of Photoshop since they're mostly used by students and hobbyists, and that increases their market share amongst professionals.

Ideas and their implementations, including software, are not private goods. Nonetheless, the current copyright system basically pretends otherwise; that supposing that something that is not nonetheless is leads to absurd results should be unsurprising.

My answer to this is similar to my answer to a lot of people's arguments about Pascal's Mugging. People tend do a lot of pseudo-utilitarian or economical analysis that leaves out the wider implications of what happens when all adopt the strategy that the author proposes as more rational. In Pascal's mugging, if everyone pays off Pascal's mugger, than everyone also knows that they can Pascal's mug anyone successfully, and they occur all the time, to which most people would assign a tremendously negative dose of utility. In the same vein, when everyone adopts the strategy of "pirate", there is little to no incentive to make the software in the first place, as TimS pointed out.